{ "best_metric": 0.05673447530806543, "best_model_checkpoint": "23_02_15_trocr_kurrent_XVI_XVIII_baseXIX_20ep/checkpoint-198500", "epoch": 19.97378880112693, "global_step": 198500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01, "learning_rate": 5e-06, "loss": 3.6945, "step": 50 }, { "epoch": 0.01, "learning_rate": 9.900000000000002e-06, "loss": 3.2928, "step": 100 }, { "epoch": 0.02, "learning_rate": 1.49e-05, "loss": 2.7794, "step": 150 }, { "epoch": 0.02, "learning_rate": 1.9900000000000003e-05, "loss": 2.238, "step": 200 }, { "epoch": 0.03, "learning_rate": 2.4900000000000002e-05, "loss": 1.9805, "step": 250 }, { "epoch": 0.03, "learning_rate": 2.9900000000000002e-05, "loss": 1.8237, "step": 300 }, { "epoch": 0.04, "learning_rate": 3.49e-05, "loss": 1.6443, "step": 350 }, { "epoch": 0.04, "learning_rate": 3.99e-05, "loss": 1.6429, "step": 400 }, { "epoch": 0.05, "learning_rate": 4.49e-05, "loss": 1.635, "step": 450 }, { "epoch": 0.05, "learning_rate": 4.99e-05, "loss": 1.5516, "step": 500 }, { "epoch": 0.05, "eval_cer": 0.217759761891419, "eval_loss": 1.4486242532730103, "eval_runtime": 1654.3825, "eval_samples_per_second": 4.929, "eval_steps_per_second": 0.617, "step": 500 }, { "epoch": 0.06, "learning_rate": 4.998764248966005e-05, "loss": 1.5053, "step": 550 }, { "epoch": 0.06, "learning_rate": 4.9975032785231515e-05, "loss": 1.4069, "step": 600 }, { "epoch": 0.07, "learning_rate": 4.996242308080299e-05, "loss": 1.4089, "step": 650 }, { "epoch": 0.07, "learning_rate": 4.994981337637446e-05, "loss": 1.3736, "step": 700 }, { "epoch": 0.08, "learning_rate": 4.993720367194593e-05, "loss": 1.3641, "step": 750 }, { "epoch": 0.08, "learning_rate": 4.99245939675174e-05, "loss": 1.3396, "step": 800 }, { "epoch": 0.09, "learning_rate": 4.9911984263088876e-05, "loss": 1.3121, "step": 850 }, { "epoch": 0.09, "learning_rate": 4.989962675274892e-05, "loss": 1.2749, "step": 900 }, { "epoch": 0.1, "learning_rate": 4.988701704832039e-05, "loss": 1.2759, "step": 950 }, { "epoch": 0.1, "learning_rate": 4.987440734389186e-05, "loss": 1.2611, "step": 1000 }, { "epoch": 0.1, "eval_cer": 0.18181308725459064, "eval_loss": 1.1947869062423706, "eval_runtime": 1431.9095, "eval_samples_per_second": 5.695, "eval_steps_per_second": 0.712, "step": 1000 }, { "epoch": 0.11, "learning_rate": 4.986179763946333e-05, "loss": 1.238, "step": 1050 }, { "epoch": 0.11, "learning_rate": 4.984918793503481e-05, "loss": 1.1745, "step": 1100 }, { "epoch": 0.12, "learning_rate": 4.983657823060628e-05, "loss": 1.2175, "step": 1150 }, { "epoch": 0.12, "learning_rate": 4.982396852617775e-05, "loss": 1.2205, "step": 1200 }, { "epoch": 0.13, "learning_rate": 4.9811358821749224e-05, "loss": 1.1743, "step": 1250 }, { "epoch": 0.13, "learning_rate": 4.979874911732069e-05, "loss": 1.1495, "step": 1300 }, { "epoch": 0.14, "learning_rate": 4.978613941289216e-05, "loss": 1.1419, "step": 1350 }, { "epoch": 0.14, "learning_rate": 4.9773529708463636e-05, "loss": 1.125, "step": 1400 }, { "epoch": 0.15, "learning_rate": 4.9760920004035104e-05, "loss": 1.1208, "step": 1450 }, { "epoch": 0.15, "learning_rate": 4.9748310299606585e-05, "loss": 1.0727, "step": 1500 }, { "epoch": 0.15, "eval_cer": 0.16596305082846094, "eval_loss": 1.0472463369369507, "eval_runtime": 1393.6622, "eval_samples_per_second": 5.851, "eval_steps_per_second": 0.732, "step": 1500 }, { "epoch": 0.16, "learning_rate": 4.973570059517805e-05, "loss": 1.1025, "step": 1550 }, { "epoch": 0.16, "learning_rate": 4.972309089074952e-05, "loss": 1.1256, "step": 1600 }, { "epoch": 0.17, "learning_rate": 4.9710481186320996e-05, "loss": 1.0788, "step": 1650 }, { "epoch": 0.17, "learning_rate": 4.9697871481892464e-05, "loss": 1.0689, "step": 1700 }, { "epoch": 0.18, "learning_rate": 4.968526177746394e-05, "loss": 1.0909, "step": 1750 }, { "epoch": 0.18, "learning_rate": 4.967265207303541e-05, "loss": 1.0333, "step": 1800 }, { "epoch": 0.19, "learning_rate": 4.966004236860688e-05, "loss": 1.0729, "step": 1850 }, { "epoch": 0.19, "learning_rate": 4.964743266417836e-05, "loss": 1.0067, "step": 1900 }, { "epoch": 0.2, "learning_rate": 4.9634822959749825e-05, "loss": 1.065, "step": 1950 }, { "epoch": 0.2, "learning_rate": 4.96222132553213e-05, "loss": 0.9611, "step": 2000 }, { "epoch": 0.2, "eval_cer": 0.15290257099981944, "eval_loss": 0.9824811220169067, "eval_runtime": 1321.3401, "eval_samples_per_second": 6.172, "eval_steps_per_second": 0.772, "step": 2000 }, { "epoch": 0.21, "learning_rate": 4.960960355089277e-05, "loss": 1.0165, "step": 2050 }, { "epoch": 0.21, "learning_rate": 4.959699384646424e-05, "loss": 1.0397, "step": 2100 }, { "epoch": 0.22, "learning_rate": 4.958438414203571e-05, "loss": 1.0071, "step": 2150 }, { "epoch": 0.22, "learning_rate": 4.957177443760718e-05, "loss": 0.9813, "step": 2200 }, { "epoch": 0.23, "learning_rate": 4.955916473317866e-05, "loss": 1.0172, "step": 2250 }, { "epoch": 0.23, "learning_rate": 4.954655502875013e-05, "loss": 0.9758, "step": 2300 }, { "epoch": 0.24, "learning_rate": 4.9533945324321604e-05, "loss": 0.9327, "step": 2350 }, { "epoch": 0.24, "learning_rate": 4.952133561989307e-05, "loss": 0.9595, "step": 2400 }, { "epoch": 0.25, "learning_rate": 4.950872591546454e-05, "loss": 1.0148, "step": 2450 }, { "epoch": 0.25, "learning_rate": 4.9496116211036015e-05, "loss": 0.9625, "step": 2500 }, { "epoch": 0.25, "eval_cer": 0.1428153350892596, "eval_loss": 0.9086866974830627, "eval_runtime": 1363.7749, "eval_samples_per_second": 5.98, "eval_steps_per_second": 0.748, "step": 2500 }, { "epoch": 0.26, "learning_rate": 4.948350650660748e-05, "loss": 1.0156, "step": 2550 }, { "epoch": 0.26, "learning_rate": 4.9470896802178965e-05, "loss": 0.9836, "step": 2600 }, { "epoch": 0.27, "learning_rate": 4.945828709775043e-05, "loss": 0.9354, "step": 2650 }, { "epoch": 0.27, "learning_rate": 4.94456773933219e-05, "loss": 0.9183, "step": 2700 }, { "epoch": 0.28, "learning_rate": 4.9433067688893376e-05, "loss": 0.9178, "step": 2750 }, { "epoch": 0.28, "learning_rate": 4.9420457984464844e-05, "loss": 0.9127, "step": 2800 }, { "epoch": 0.29, "learning_rate": 4.940784828003632e-05, "loss": 0.9183, "step": 2850 }, { "epoch": 0.29, "learning_rate": 4.939523857560779e-05, "loss": 0.8711, "step": 2900 }, { "epoch": 0.3, "learning_rate": 4.938262887117926e-05, "loss": 0.9306, "step": 2950 }, { "epoch": 0.3, "learning_rate": 4.9370019166750737e-05, "loss": 0.8859, "step": 3000 }, { "epoch": 0.3, "eval_cer": 0.1392505557319784, "eval_loss": 0.8707912564277649, "eval_runtime": 1427.3832, "eval_samples_per_second": 5.713, "eval_steps_per_second": 0.715, "step": 3000 }, { "epoch": 0.31, "learning_rate": 4.9357409462322205e-05, "loss": 0.9036, "step": 3050 }, { "epoch": 0.31, "learning_rate": 4.934479975789368e-05, "loss": 0.9136, "step": 3100 }, { "epoch": 0.32, "learning_rate": 4.933219005346515e-05, "loss": 0.8985, "step": 3150 }, { "epoch": 0.32, "learning_rate": 4.931983254312519e-05, "loss": 0.868, "step": 3200 }, { "epoch": 0.33, "learning_rate": 4.930722283869667e-05, "loss": 0.9057, "step": 3250 }, { "epoch": 0.33, "learning_rate": 4.9294613134268135e-05, "loss": 0.839, "step": 3300 }, { "epoch": 0.34, "learning_rate": 4.928200342983961e-05, "loss": 0.8637, "step": 3350 }, { "epoch": 0.34, "learning_rate": 4.926939372541108e-05, "loss": 0.8906, "step": 3400 }, { "epoch": 0.35, "learning_rate": 4.9256784020982546e-05, "loss": 0.9164, "step": 3450 }, { "epoch": 0.35, "learning_rate": 4.924417431655402e-05, "loss": 0.844, "step": 3500 }, { "epoch": 0.35, "eval_cer": 0.136078057771219, "eval_loss": 0.8419156074523926, "eval_runtime": 1641.553, "eval_samples_per_second": 4.968, "eval_steps_per_second": 0.621, "step": 3500 }, { "epoch": 0.36, "learning_rate": 4.923156461212549e-05, "loss": 0.8291, "step": 3550 }, { "epoch": 0.36, "learning_rate": 4.921895490769697e-05, "loss": 0.8979, "step": 3600 }, { "epoch": 0.37, "learning_rate": 4.920634520326844e-05, "loss": 0.8776, "step": 3650 }, { "epoch": 0.37, "learning_rate": 4.919373549883991e-05, "loss": 0.8312, "step": 3700 }, { "epoch": 0.38, "learning_rate": 4.918112579441138e-05, "loss": 0.8367, "step": 3750 }, { "epoch": 0.38, "learning_rate": 4.916851608998285e-05, "loss": 0.8153, "step": 3800 }, { "epoch": 0.39, "learning_rate": 4.9155906385554325e-05, "loss": 0.8236, "step": 3850 }, { "epoch": 0.39, "learning_rate": 4.914329668112579e-05, "loss": 0.8227, "step": 3900 }, { "epoch": 0.4, "learning_rate": 4.913068697669727e-05, "loss": 0.8299, "step": 3950 }, { "epoch": 0.4, "learning_rate": 4.911807727226874e-05, "loss": 0.7874, "step": 4000 }, { "epoch": 0.4, "eval_cer": 0.1287243382586442, "eval_loss": 0.7957796454429626, "eval_runtime": 1672.3091, "eval_samples_per_second": 4.876, "eval_steps_per_second": 0.61, "step": 4000 }, { "epoch": 0.41, "learning_rate": 4.910546756784021e-05, "loss": 0.8161, "step": 4050 }, { "epoch": 0.41, "learning_rate": 4.9092857863411686e-05, "loss": 0.825, "step": 4100 }, { "epoch": 0.42, "learning_rate": 4.9080248158983154e-05, "loss": 0.8088, "step": 4150 }, { "epoch": 0.42, "learning_rate": 4.906763845455463e-05, "loss": 0.8207, "step": 4200 }, { "epoch": 0.43, "learning_rate": 4.90550287501261e-05, "loss": 0.7728, "step": 4250 }, { "epoch": 0.43, "learning_rate": 4.904241904569757e-05, "loss": 0.8668, "step": 4300 }, { "epoch": 0.44, "learning_rate": 4.9029809341269047e-05, "loss": 0.7698, "step": 4350 }, { "epoch": 0.44, "learning_rate": 4.9017199636840515e-05, "loss": 0.7996, "step": 4400 }, { "epoch": 0.45, "learning_rate": 4.900458993241199e-05, "loss": 0.8034, "step": 4450 }, { "epoch": 0.45, "learning_rate": 4.899198022798346e-05, "loss": 0.7818, "step": 4500 }, { "epoch": 0.45, "eval_cer": 0.1300879831132199, "eval_loss": 0.7830387353897095, "eval_runtime": 1611.973, "eval_samples_per_second": 5.059, "eval_steps_per_second": 0.633, "step": 4500 }, { "epoch": 0.46, "learning_rate": 4.897937052355493e-05, "loss": 0.7718, "step": 4550 }, { "epoch": 0.46, "learning_rate": 4.89667608191264e-05, "loss": 0.8315, "step": 4600 }, { "epoch": 0.47, "learning_rate": 4.895415111469787e-05, "loss": 0.8133, "step": 4650 }, { "epoch": 0.47, "learning_rate": 4.894154141026935e-05, "loss": 0.8126, "step": 4700 }, { "epoch": 0.48, "learning_rate": 4.892893170584082e-05, "loss": 0.7871, "step": 4750 }, { "epoch": 0.48, "learning_rate": 4.891632200141229e-05, "loss": 0.7632, "step": 4800 }, { "epoch": 0.49, "learning_rate": 4.890371229698376e-05, "loss": 0.7989, "step": 4850 }, { "epoch": 0.49, "learning_rate": 4.889110259255523e-05, "loss": 0.7792, "step": 4900 }, { "epoch": 0.5, "learning_rate": 4.8878492888126704e-05, "loss": 0.7837, "step": 4950 }, { "epoch": 0.5, "learning_rate": 4.886588318369817e-05, "loss": 0.7687, "step": 5000 }, { "epoch": 0.5, "eval_cer": 0.12189366060809843, "eval_loss": 0.7457525134086609, "eval_runtime": 1620.4632, "eval_samples_per_second": 5.033, "eval_steps_per_second": 0.629, "step": 5000 }, { "epoch": 0.51, "learning_rate": 4.885327347926965e-05, "loss": 0.7659, "step": 5050 }, { "epoch": 0.51, "learning_rate": 4.884066377484112e-05, "loss": 0.7799, "step": 5100 }, { "epoch": 0.52, "learning_rate": 4.882805407041259e-05, "loss": 0.7525, "step": 5150 }, { "epoch": 0.52, "learning_rate": 4.8815444365984065e-05, "loss": 0.7694, "step": 5200 }, { "epoch": 0.53, "learning_rate": 4.880283466155553e-05, "loss": 0.77, "step": 5250 }, { "epoch": 0.53, "learning_rate": 4.879022495712701e-05, "loss": 0.7518, "step": 5300 }, { "epoch": 0.54, "learning_rate": 4.8777615252698476e-05, "loss": 0.7647, "step": 5350 }, { "epoch": 0.54, "learning_rate": 4.876525774235852e-05, "loss": 0.7863, "step": 5400 }, { "epoch": 0.55, "learning_rate": 4.8752648037929996e-05, "loss": 0.8033, "step": 5450 }, { "epoch": 0.55, "learning_rate": 4.8740038333501464e-05, "loss": 0.7596, "step": 5500 }, { "epoch": 0.55, "eval_cer": 0.11943411851879526, "eval_loss": 0.7319559454917908, "eval_runtime": 1373.4733, "eval_samples_per_second": 5.938, "eval_steps_per_second": 0.743, "step": 5500 }, { "epoch": 0.56, "learning_rate": 4.872742862907293e-05, "loss": 0.727, "step": 5550 }, { "epoch": 0.56, "learning_rate": 4.871481892464441e-05, "loss": 0.7723, "step": 5600 }, { "epoch": 0.57, "learning_rate": 4.870220922021588e-05, "loss": 0.7298, "step": 5650 }, { "epoch": 0.57, "learning_rate": 4.868959951578736e-05, "loss": 0.7809, "step": 5700 }, { "epoch": 0.58, "learning_rate": 4.8676989811358825e-05, "loss": 0.7043, "step": 5750 }, { "epoch": 0.58, "learning_rate": 4.86643801069303e-05, "loss": 0.7472, "step": 5800 }, { "epoch": 0.59, "learning_rate": 4.865177040250177e-05, "loss": 0.7694, "step": 5850 }, { "epoch": 0.59, "learning_rate": 4.8639160698073236e-05, "loss": 0.7269, "step": 5900 }, { "epoch": 0.6, "learning_rate": 4.862655099364471e-05, "loss": 0.7317, "step": 5950 }, { "epoch": 0.6, "learning_rate": 4.861394128921618e-05, "loss": 0.7637, "step": 6000 }, { "epoch": 0.6, "eval_cer": 0.11985130667065175, "eval_loss": 0.7123238444328308, "eval_runtime": 1348.9228, "eval_samples_per_second": 6.046, "eval_steps_per_second": 0.756, "step": 6000 }, { "epoch": 0.61, "learning_rate": 4.860133158478766e-05, "loss": 0.7322, "step": 6050 }, { "epoch": 0.61, "learning_rate": 4.858872188035913e-05, "loss": 0.7217, "step": 6100 }, { "epoch": 0.62, "learning_rate": 4.8576112175930597e-05, "loss": 0.7495, "step": 6150 }, { "epoch": 0.62, "learning_rate": 4.856350247150207e-05, "loss": 0.7316, "step": 6200 }, { "epoch": 0.63, "learning_rate": 4.855089276707354e-05, "loss": 0.7491, "step": 6250 }, { "epoch": 0.63, "learning_rate": 4.8538283062645014e-05, "loss": 0.7318, "step": 6300 }, { "epoch": 0.64, "learning_rate": 4.852567335821648e-05, "loss": 0.7403, "step": 6350 }, { "epoch": 0.64, "learning_rate": 4.851306365378796e-05, "loss": 0.7226, "step": 6400 }, { "epoch": 0.65, "learning_rate": 4.850045394935943e-05, "loss": 0.7047, "step": 6450 }, { "epoch": 0.65, "learning_rate": 4.84878442449309e-05, "loss": 0.7471, "step": 6500 }, { "epoch": 0.65, "eval_cer": 0.1135903710483876, "eval_loss": 0.6905709505081177, "eval_runtime": 1342.0798, "eval_samples_per_second": 6.076, "eval_steps_per_second": 0.76, "step": 6500 }, { "epoch": 0.66, "learning_rate": 4.8475234540502375e-05, "loss": 0.7044, "step": 6550 }, { "epoch": 0.66, "learning_rate": 4.846262483607384e-05, "loss": 0.7117, "step": 6600 }, { "epoch": 0.67, "learning_rate": 4.845001513164532e-05, "loss": 0.7083, "step": 6650 }, { "epoch": 0.67, "learning_rate": 4.8437405427216786e-05, "loss": 0.7209, "step": 6700 }, { "epoch": 0.68, "learning_rate": 4.8424795722788254e-05, "loss": 0.7247, "step": 6750 }, { "epoch": 0.68, "learning_rate": 4.8412186018359736e-05, "loss": 0.7032, "step": 6800 }, { "epoch": 0.69, "learning_rate": 4.8399576313931204e-05, "loss": 0.7347, "step": 6850 }, { "epoch": 0.69, "learning_rate": 4.838696660950268e-05, "loss": 0.6771, "step": 6900 }, { "epoch": 0.7, "learning_rate": 4.837435690507415e-05, "loss": 0.723, "step": 6950 }, { "epoch": 0.7, "learning_rate": 4.8361747200645615e-05, "loss": 0.6581, "step": 7000 }, { "epoch": 0.7, "eval_cer": 0.11170057098736605, "eval_loss": 0.6762269735336304, "eval_runtime": 1423.0043, "eval_samples_per_second": 5.731, "eval_steps_per_second": 0.717, "step": 7000 }, { "epoch": 0.71, "learning_rate": 4.834913749621709e-05, "loss": 0.7186, "step": 7050 }, { "epoch": 0.71, "learning_rate": 4.833652779178856e-05, "loss": 0.7546, "step": 7100 }, { "epoch": 0.72, "learning_rate": 4.832391808736004e-05, "loss": 0.6776, "step": 7150 }, { "epoch": 0.72, "learning_rate": 4.831130838293151e-05, "loss": 0.675, "step": 7200 }, { "epoch": 0.73, "learning_rate": 4.8298698678502976e-05, "loss": 0.6826, "step": 7250 }, { "epoch": 0.73, "learning_rate": 4.828608897407445e-05, "loss": 0.7097, "step": 7300 }, { "epoch": 0.74, "learning_rate": 4.827347926964592e-05, "loss": 0.7244, "step": 7350 }, { "epoch": 0.74, "learning_rate": 4.8260869565217394e-05, "loss": 0.6731, "step": 7400 }, { "epoch": 0.75, "learning_rate": 4.824825986078886e-05, "loss": 0.691, "step": 7450 }, { "epoch": 0.75, "learning_rate": 4.823565015636034e-05, "loss": 0.6858, "step": 7500 }, { "epoch": 0.75, "eval_cer": 0.11321054302953318, "eval_loss": 0.6656007170677185, "eval_runtime": 1729.7029, "eval_samples_per_second": 4.715, "eval_steps_per_second": 0.59, "step": 7500 }, { "epoch": 0.76, "learning_rate": 4.822304045193181e-05, "loss": 0.7071, "step": 7550 }, { "epoch": 0.76, "learning_rate": 4.821068294159185e-05, "loss": 0.6989, "step": 7600 }, { "epoch": 0.77, "learning_rate": 4.8198073237163324e-05, "loss": 0.6794, "step": 7650 }, { "epoch": 0.77, "learning_rate": 4.818546353273479e-05, "loss": 0.6471, "step": 7700 }, { "epoch": 0.78, "learning_rate": 4.817285382830627e-05, "loss": 0.6466, "step": 7750 }, { "epoch": 0.78, "learning_rate": 4.816024412387774e-05, "loss": 0.6798, "step": 7800 }, { "epoch": 0.79, "learning_rate": 4.814763441944921e-05, "loss": 0.6996, "step": 7850 }, { "epoch": 0.79, "learning_rate": 4.8135024715020685e-05, "loss": 0.6875, "step": 7900 }, { "epoch": 0.8, "learning_rate": 4.812241501059215e-05, "loss": 0.6843, "step": 7950 }, { "epoch": 0.8, "learning_rate": 4.810980530616362e-05, "loss": 0.6942, "step": 8000 }, { "epoch": 0.8, "eval_cer": 0.11137989651243159, "eval_loss": 0.6471945643424988, "eval_runtime": 1675.2325, "eval_samples_per_second": 4.868, "eval_steps_per_second": 0.609, "step": 8000 }, { "epoch": 0.81, "learning_rate": 4.8097195601735096e-05, "loss": 0.6479, "step": 8050 }, { "epoch": 0.82, "learning_rate": 4.808458589730657e-05, "loss": 0.7053, "step": 8100 }, { "epoch": 0.82, "learning_rate": 4.8071976192878046e-05, "loss": 0.685, "step": 8150 }, { "epoch": 0.83, "learning_rate": 4.8059366488449514e-05, "loss": 0.6518, "step": 8200 }, { "epoch": 0.83, "learning_rate": 4.804675678402098e-05, "loss": 0.668, "step": 8250 }, { "epoch": 0.84, "learning_rate": 4.803414707959246e-05, "loss": 0.7153, "step": 8300 }, { "epoch": 0.84, "learning_rate": 4.8021537375163925e-05, "loss": 0.6547, "step": 8350 }, { "epoch": 0.85, "learning_rate": 4.80089276707354e-05, "loss": 0.6549, "step": 8400 }, { "epoch": 0.85, "learning_rate": 4.799631796630687e-05, "loss": 0.6607, "step": 8450 }, { "epoch": 0.86, "learning_rate": 4.798370826187834e-05, "loss": 0.6656, "step": 8500 }, { "epoch": 0.86, "eval_cer": 0.10565134278544698, "eval_loss": 0.6389959454536438, "eval_runtime": 1528.9697, "eval_samples_per_second": 5.334, "eval_steps_per_second": 0.667, "step": 8500 }, { "epoch": 0.86, "learning_rate": 4.797109855744982e-05, "loss": 0.6453, "step": 8550 }, { "epoch": 0.87, "learning_rate": 4.7958488853021286e-05, "loss": 0.6667, "step": 8600 }, { "epoch": 0.87, "learning_rate": 4.794587914859276e-05, "loss": 0.6381, "step": 8650 }, { "epoch": 0.88, "learning_rate": 4.793326944416423e-05, "loss": 0.6776, "step": 8700 }, { "epoch": 0.88, "learning_rate": 4.7920659739735704e-05, "loss": 0.6232, "step": 8750 }, { "epoch": 0.89, "learning_rate": 4.790805003530717e-05, "loss": 0.6822, "step": 8800 }, { "epoch": 0.89, "learning_rate": 4.789544033087865e-05, "loss": 0.6648, "step": 8850 }, { "epoch": 0.9, "learning_rate": 4.788283062645012e-05, "loss": 0.6595, "step": 8900 }, { "epoch": 0.9, "learning_rate": 4.787022092202159e-05, "loss": 0.6522, "step": 8950 }, { "epoch": 0.91, "learning_rate": 4.7857611217593065e-05, "loss": 0.6479, "step": 9000 }, { "epoch": 0.91, "eval_cer": 0.10444959184054695, "eval_loss": 0.6251552104949951, "eval_runtime": 1400.9777, "eval_samples_per_second": 5.821, "eval_steps_per_second": 0.728, "step": 9000 }, { "epoch": 0.91, "learning_rate": 4.784500151316453e-05, "loss": 0.6618, "step": 9050 }, { "epoch": 0.92, "learning_rate": 4.783239180873601e-05, "loss": 0.6929, "step": 9100 }, { "epoch": 0.92, "learning_rate": 4.7819782104307476e-05, "loss": 0.6244, "step": 9150 }, { "epoch": 0.93, "learning_rate": 4.7807172399878944e-05, "loss": 0.6757, "step": 9200 }, { "epoch": 0.93, "learning_rate": 4.7794562695450426e-05, "loss": 0.6166, "step": 9250 }, { "epoch": 0.94, "learning_rate": 4.7781952991021894e-05, "loss": 0.6728, "step": 9300 }, { "epoch": 0.94, "learning_rate": 4.776934328659337e-05, "loss": 0.6207, "step": 9350 }, { "epoch": 0.95, "learning_rate": 4.7756733582164837e-05, "loss": 0.6533, "step": 9400 }, { "epoch": 0.95, "learning_rate": 4.7744123877736305e-05, "loss": 0.6654, "step": 9450 }, { "epoch": 0.96, "learning_rate": 4.773151417330778e-05, "loss": 0.6491, "step": 9500 }, { "epoch": 0.96, "eval_cer": 0.1040822171993599, "eval_loss": 0.614883542060852, "eval_runtime": 1665.1117, "eval_samples_per_second": 4.898, "eval_steps_per_second": 0.613, "step": 9500 }, { "epoch": 0.96, "learning_rate": 4.771890446887925e-05, "loss": 0.6607, "step": 9550 }, { "epoch": 0.97, "learning_rate": 4.770629476445073e-05, "loss": 0.6298, "step": 9600 }, { "epoch": 0.97, "learning_rate": 4.76936850600222e-05, "loss": 0.652, "step": 9650 }, { "epoch": 0.98, "learning_rate": 4.7681075355593665e-05, "loss": 0.6451, "step": 9700 }, { "epoch": 0.98, "learning_rate": 4.766846565116514e-05, "loss": 0.6097, "step": 9750 }, { "epoch": 0.99, "learning_rate": 4.765610814082518e-05, "loss": 0.6178, "step": 9800 }, { "epoch": 0.99, "learning_rate": 4.764349843639665e-05, "loss": 0.6431, "step": 9850 }, { "epoch": 1.0, "learning_rate": 4.763088873196813e-05, "loss": 0.6367, "step": 9900 }, { "epoch": 1.0, "learning_rate": 4.7618279027539596e-05, "loss": 0.5976, "step": 9950 }, { "epoch": 1.01, "learning_rate": 4.760566932311107e-05, "loss": 0.5021, "step": 10000 }, { "epoch": 1.01, "eval_cer": 0.10476403962664774, "eval_loss": 0.6128339767456055, "eval_runtime": 1504.159, "eval_samples_per_second": 5.422, "eval_steps_per_second": 0.678, "step": 10000 }, { "epoch": 1.01, "learning_rate": 4.759305961868254e-05, "loss": 0.4951, "step": 10050 }, { "epoch": 1.02, "learning_rate": 4.758044991425401e-05, "loss": 0.5154, "step": 10100 }, { "epoch": 1.02, "learning_rate": 4.756784020982548e-05, "loss": 0.5094, "step": 10150 }, { "epoch": 1.03, "learning_rate": 4.755523050539696e-05, "loss": 0.5004, "step": 10200 }, { "epoch": 1.03, "learning_rate": 4.754262080096843e-05, "loss": 0.5164, "step": 10250 }, { "epoch": 1.04, "learning_rate": 4.75300110965399e-05, "loss": 0.4835, "step": 10300 }, { "epoch": 1.04, "learning_rate": 4.7517401392111375e-05, "loss": 0.5167, "step": 10350 }, { "epoch": 1.05, "learning_rate": 4.750479168768284e-05, "loss": 0.5084, "step": 10400 }, { "epoch": 1.05, "learning_rate": 4.749218198325431e-05, "loss": 0.5103, "step": 10450 }, { "epoch": 1.06, "learning_rate": 4.7479572278825786e-05, "loss": 0.5111, "step": 10500 }, { "epoch": 1.06, "eval_cer": 0.09966438147186471, "eval_loss": 0.6031123995780945, "eval_runtime": 1679.3868, "eval_samples_per_second": 4.856, "eval_steps_per_second": 0.607, "step": 10500 }, { "epoch": 1.06, "learning_rate": 4.7466962574397254e-05, "loss": 0.5152, "step": 10550 }, { "epoch": 1.07, "learning_rate": 4.7454352869968736e-05, "loss": 0.5019, "step": 10600 }, { "epoch": 1.07, "learning_rate": 4.7441743165540204e-05, "loss": 0.505, "step": 10650 }, { "epoch": 1.08, "learning_rate": 4.742913346111167e-05, "loss": 0.5044, "step": 10700 }, { "epoch": 1.08, "learning_rate": 4.741652375668315e-05, "loss": 0.52, "step": 10750 }, { "epoch": 1.09, "learning_rate": 4.7403914052254615e-05, "loss": 0.4945, "step": 10800 }, { "epoch": 1.09, "learning_rate": 4.739130434782609e-05, "loss": 0.5214, "step": 10850 }, { "epoch": 1.1, "learning_rate": 4.737869464339756e-05, "loss": 0.523, "step": 10900 }, { "epoch": 1.1, "learning_rate": 4.736608493896903e-05, "loss": 0.4901, "step": 10950 }, { "epoch": 1.11, "learning_rate": 4.735347523454051e-05, "loss": 0.4959, "step": 11000 }, { "epoch": 1.11, "eval_cer": 0.10257535850160959, "eval_loss": 0.6009875535964966, "eval_runtime": 1762.8055, "eval_samples_per_second": 4.626, "eval_steps_per_second": 0.579, "step": 11000 }, { "epoch": 1.11, "learning_rate": 4.7340865530111976e-05, "loss": 0.513, "step": 11050 }, { "epoch": 1.12, "learning_rate": 4.732825582568345e-05, "loss": 0.5102, "step": 11100 }, { "epoch": 1.12, "learning_rate": 4.731564612125492e-05, "loss": 0.5328, "step": 11150 }, { "epoch": 1.13, "learning_rate": 4.730303641682639e-05, "loss": 0.5099, "step": 11200 }, { "epoch": 1.13, "learning_rate": 4.729042671239786e-05, "loss": 0.4553, "step": 11250 }, { "epoch": 1.14, "learning_rate": 4.727781700796933e-05, "loss": 0.5407, "step": 11300 }, { "epoch": 1.14, "learning_rate": 4.726520730354081e-05, "loss": 0.5155, "step": 11350 }, { "epoch": 1.15, "learning_rate": 4.725259759911228e-05, "loss": 0.461, "step": 11400 }, { "epoch": 1.15, "learning_rate": 4.7239987894683754e-05, "loss": 0.4957, "step": 11450 }, { "epoch": 1.16, "learning_rate": 4.722737819025522e-05, "loss": 0.5299, "step": 11500 }, { "epoch": 1.16, "eval_cer": 0.10072291857359013, "eval_loss": 0.5908301472663879, "eval_runtime": 1774.0841, "eval_samples_per_second": 4.597, "eval_steps_per_second": 0.575, "step": 11500 }, { "epoch": 1.16, "learning_rate": 4.721476848582669e-05, "loss": 0.4635, "step": 11550 }, { "epoch": 1.17, "learning_rate": 4.7202158781398165e-05, "loss": 0.5083, "step": 11600 }, { "epoch": 1.17, "learning_rate": 4.718954907696963e-05, "loss": 0.5131, "step": 11650 }, { "epoch": 1.18, "learning_rate": 4.7176939372541115e-05, "loss": 0.5215, "step": 11700 }, { "epoch": 1.18, "learning_rate": 4.716432966811258e-05, "loss": 0.5086, "step": 11750 }, { "epoch": 1.19, "learning_rate": 4.715171996368405e-05, "loss": 0.5114, "step": 11800 }, { "epoch": 1.19, "learning_rate": 4.7139110259255526e-05, "loss": 0.5221, "step": 11850 }, { "epoch": 1.2, "learning_rate": 4.7126500554826994e-05, "loss": 0.4625, "step": 11900 }, { "epoch": 1.2, "learning_rate": 4.711389085039847e-05, "loss": 0.4998, "step": 11950 }, { "epoch": 1.21, "learning_rate": 4.710128114596994e-05, "loss": 0.4951, "step": 12000 }, { "epoch": 1.21, "eval_cer": 0.09878330500189914, "eval_loss": 0.5946286916732788, "eval_runtime": 1478.8431, "eval_samples_per_second": 5.514, "eval_steps_per_second": 0.69, "step": 12000 }, { "epoch": 1.21, "learning_rate": 4.708867144154141e-05, "loss": 0.4931, "step": 12050 }, { "epoch": 1.22, "learning_rate": 4.707606173711289e-05, "loss": 0.5315, "step": 12100 }, { "epoch": 1.22, "learning_rate": 4.7063452032684355e-05, "loss": 0.5054, "step": 12150 }, { "epoch": 1.23, "learning_rate": 4.705084232825583e-05, "loss": 0.4964, "step": 12200 }, { "epoch": 1.23, "learning_rate": 4.70382326238273e-05, "loss": 0.5245, "step": 12250 }, { "epoch": 1.24, "learning_rate": 4.702562291939877e-05, "loss": 0.4866, "step": 12300 }, { "epoch": 1.24, "learning_rate": 4.701301321497024e-05, "loss": 0.5146, "step": 12350 }, { "epoch": 1.25, "learning_rate": 4.7000403510541716e-05, "loss": 0.5082, "step": 12400 }, { "epoch": 1.25, "learning_rate": 4.698779380611319e-05, "loss": 0.521, "step": 12450 }, { "epoch": 1.26, "learning_rate": 4.697543629577323e-05, "loss": 0.4751, "step": 12500 }, { "epoch": 1.26, "eval_cer": 0.09834432343912478, "eval_loss": 0.582065224647522, "eval_runtime": 1483.4631, "eval_samples_per_second": 5.497, "eval_steps_per_second": 0.688, "step": 12500 }, { "epoch": 1.26, "learning_rate": 4.69628265913447e-05, "loss": 0.4988, "step": 12550 }, { "epoch": 1.27, "learning_rate": 4.695021688691617e-05, "loss": 0.4517, "step": 12600 }, { "epoch": 1.27, "learning_rate": 4.6937607182487646e-05, "loss": 0.493, "step": 12650 }, { "epoch": 1.28, "learning_rate": 4.692499747805912e-05, "loss": 0.4883, "step": 12700 }, { "epoch": 1.28, "learning_rate": 4.691238777363059e-05, "loss": 0.4947, "step": 12750 }, { "epoch": 1.29, "learning_rate": 4.689977806920206e-05, "loss": 0.4948, "step": 12800 }, { "epoch": 1.29, "learning_rate": 4.688716836477353e-05, "loss": 0.4986, "step": 12850 }, { "epoch": 1.3, "learning_rate": 4.6874558660345e-05, "loss": 0.5058, "step": 12900 }, { "epoch": 1.3, "learning_rate": 4.6861948955916475e-05, "loss": 0.4782, "step": 12950 }, { "epoch": 1.31, "learning_rate": 4.684933925148794e-05, "loss": 0.5078, "step": 13000 }, { "epoch": 1.31, "eval_cer": 0.09526833915528739, "eval_loss": 0.5764759182929993, "eval_runtime": 1312.8817, "eval_samples_per_second": 6.212, "eval_steps_per_second": 0.777, "step": 13000 }, { "epoch": 1.31, "learning_rate": 4.683672954705942e-05, "loss": 0.5033, "step": 13050 }, { "epoch": 1.32, "learning_rate": 4.682411984263089e-05, "loss": 0.4908, "step": 13100 }, { "epoch": 1.32, "learning_rate": 4.681151013820236e-05, "loss": 0.4885, "step": 13150 }, { "epoch": 1.33, "learning_rate": 4.6798900433773836e-05, "loss": 0.5088, "step": 13200 }, { "epoch": 1.33, "learning_rate": 4.6786290729345304e-05, "loss": 0.4849, "step": 13250 }, { "epoch": 1.34, "learning_rate": 4.677368102491678e-05, "loss": 0.4661, "step": 13300 }, { "epoch": 1.34, "learning_rate": 4.676107132048825e-05, "loss": 0.4827, "step": 13350 }, { "epoch": 1.35, "learning_rate": 4.674846161605972e-05, "loss": 0.506, "step": 13400 }, { "epoch": 1.35, "learning_rate": 4.67358519116312e-05, "loss": 0.5037, "step": 13450 }, { "epoch": 1.36, "learning_rate": 4.6723242207202665e-05, "loss": 0.5204, "step": 13500 }, { "epoch": 1.36, "eval_cer": 0.09984806879245824, "eval_loss": 0.5712337493896484, "eval_runtime": 1587.3195, "eval_samples_per_second": 5.138, "eval_steps_per_second": 0.643, "step": 13500 }, { "epoch": 1.36, "learning_rate": 4.671063250277414e-05, "loss": 0.5424, "step": 13550 }, { "epoch": 1.37, "learning_rate": 4.669802279834561e-05, "loss": 0.4849, "step": 13600 }, { "epoch": 1.37, "learning_rate": 4.668541309391708e-05, "loss": 0.4855, "step": 13650 }, { "epoch": 1.38, "learning_rate": 4.667280338948855e-05, "loss": 0.475, "step": 13700 }, { "epoch": 1.38, "learning_rate": 4.666019368506002e-05, "loss": 0.5271, "step": 13750 }, { "epoch": 1.39, "learning_rate": 4.66475839806315e-05, "loss": 0.4718, "step": 13800 }, { "epoch": 1.39, "learning_rate": 4.663497427620297e-05, "loss": 0.4838, "step": 13850 }, { "epoch": 1.4, "learning_rate": 4.6622364571774444e-05, "loss": 0.4763, "step": 13900 }, { "epoch": 1.4, "learning_rate": 4.660975486734591e-05, "loss": 0.489, "step": 13950 }, { "epoch": 1.41, "learning_rate": 4.659714516291738e-05, "loss": 0.4878, "step": 14000 }, { "epoch": 1.41, "eval_cer": 0.09421291539797881, "eval_loss": 0.5720551013946533, "eval_runtime": 1691.3925, "eval_samples_per_second": 4.821, "eval_steps_per_second": 0.603, "step": 14000 }, { "epoch": 1.41, "learning_rate": 4.6584535458488855e-05, "loss": 0.4636, "step": 14050 }, { "epoch": 1.42, "learning_rate": 4.657192575406032e-05, "loss": 0.5162, "step": 14100 }, { "epoch": 1.42, "learning_rate": 4.6559316049631804e-05, "loss": 0.5007, "step": 14150 }, { "epoch": 1.43, "learning_rate": 4.654670634520327e-05, "loss": 0.4815, "step": 14200 }, { "epoch": 1.43, "learning_rate": 4.653409664077474e-05, "loss": 0.5147, "step": 14250 }, { "epoch": 1.44, "learning_rate": 4.6521486936346215e-05, "loss": 0.4863, "step": 14300 }, { "epoch": 1.44, "learning_rate": 4.6508877231917684e-05, "loss": 0.4936, "step": 14350 }, { "epoch": 1.45, "learning_rate": 4.649626752748916e-05, "loss": 0.4944, "step": 14400 }, { "epoch": 1.45, "learning_rate": 4.6483657823060627e-05, "loss": 0.5139, "step": 14450 }, { "epoch": 1.46, "learning_rate": 4.64710481186321e-05, "loss": 0.4597, "step": 14500 }, { "epoch": 1.46, "eval_cer": 0.09367430681386558, "eval_loss": 0.5658307671546936, "eval_runtime": 1641.4323, "eval_samples_per_second": 4.968, "eval_steps_per_second": 0.621, "step": 14500 }, { "epoch": 1.46, "learning_rate": 4.6458438414203576e-05, "loss": 0.5067, "step": 14550 }, { "epoch": 1.47, "learning_rate": 4.6445828709775044e-05, "loss": 0.4737, "step": 14600 }, { "epoch": 1.47, "learning_rate": 4.643321900534652e-05, "loss": 0.5076, "step": 14650 }, { "epoch": 1.48, "learning_rate": 4.642060930091799e-05, "loss": 0.5039, "step": 14700 }, { "epoch": 1.48, "learning_rate": 4.640799959648946e-05, "loss": 0.4689, "step": 14750 }, { "epoch": 1.49, "learning_rate": 4.639538989206093e-05, "loss": 0.4973, "step": 14800 }, { "epoch": 1.49, "learning_rate": 4.6382780187632405e-05, "loss": 0.4814, "step": 14850 }, { "epoch": 1.5, "learning_rate": 4.637017048320388e-05, "loss": 0.4936, "step": 14900 }, { "epoch": 1.5, "learning_rate": 4.635756077877535e-05, "loss": 0.4902, "step": 14950 }, { "epoch": 1.51, "learning_rate": 4.634495107434682e-05, "loss": 0.4512, "step": 15000 }, { "epoch": 1.51, "eval_cer": 0.09378016052403813, "eval_loss": 0.5478760600090027, "eval_runtime": 1746.1745, "eval_samples_per_second": 4.67, "eval_steps_per_second": 0.584, "step": 15000 }, { "epoch": 1.51, "learning_rate": 4.633259356400686e-05, "loss": 0.4579, "step": 15050 }, { "epoch": 1.52, "learning_rate": 4.6319983859578336e-05, "loss": 0.482, "step": 15100 }, { "epoch": 1.52, "learning_rate": 4.630737415514981e-05, "loss": 0.509, "step": 15150 }, { "epoch": 1.53, "learning_rate": 4.629476445072128e-05, "loss": 0.4603, "step": 15200 }, { "epoch": 1.53, "learning_rate": 4.628215474629275e-05, "loss": 0.4707, "step": 15250 }, { "epoch": 1.54, "learning_rate": 4.626954504186422e-05, "loss": 0.499, "step": 15300 }, { "epoch": 1.54, "learning_rate": 4.625693533743569e-05, "loss": 0.4902, "step": 15350 }, { "epoch": 1.55, "learning_rate": 4.6244325633007165e-05, "loss": 0.4694, "step": 15400 }, { "epoch": 1.55, "learning_rate": 4.623171592857863e-05, "loss": 0.5066, "step": 15450 }, { "epoch": 1.56, "learning_rate": 4.621910622415011e-05, "loss": 0.4482, "step": 15500 }, { "epoch": 1.56, "eval_cer": 0.09168176638708833, "eval_loss": 0.5477092862129211, "eval_runtime": 1656.9567, "eval_samples_per_second": 4.922, "eval_steps_per_second": 0.616, "step": 15500 }, { "epoch": 1.56, "learning_rate": 4.620649651972158e-05, "loss": 0.4738, "step": 15550 }, { "epoch": 1.57, "learning_rate": 4.619388681529305e-05, "loss": 0.5004, "step": 15600 }, { "epoch": 1.57, "learning_rate": 4.6181277110864526e-05, "loss": 0.5142, "step": 15650 }, { "epoch": 1.58, "learning_rate": 4.6168667406435994e-05, "loss": 0.4867, "step": 15700 }, { "epoch": 1.58, "learning_rate": 4.615605770200747e-05, "loss": 0.4579, "step": 15750 }, { "epoch": 1.59, "learning_rate": 4.6143447997578937e-05, "loss": 0.4698, "step": 15800 }, { "epoch": 1.59, "learning_rate": 4.613083829315041e-05, "loss": 0.4993, "step": 15850 }, { "epoch": 1.6, "learning_rate": 4.6118228588721886e-05, "loss": 0.4713, "step": 15900 }, { "epoch": 1.6, "learning_rate": 4.6105871078381924e-05, "loss": 0.4818, "step": 15950 }, { "epoch": 1.61, "learning_rate": 4.609326137395339e-05, "loss": 0.4923, "step": 16000 }, { "epoch": 1.61, "eval_cer": 0.0934376926381858, "eval_loss": 0.543571412563324, "eval_runtime": 1379.8992, "eval_samples_per_second": 5.91, "eval_steps_per_second": 0.739, "step": 16000 }, { "epoch": 1.61, "learning_rate": 4.608065166952487e-05, "loss": 0.5071, "step": 16050 }, { "epoch": 1.62, "learning_rate": 4.606804196509634e-05, "loss": 0.4848, "step": 16100 }, { "epoch": 1.63, "learning_rate": 4.605543226066781e-05, "loss": 0.4761, "step": 16150 }, { "epoch": 1.63, "learning_rate": 4.6042822556239285e-05, "loss": 0.4523, "step": 16200 }, { "epoch": 1.64, "learning_rate": 4.603021285181075e-05, "loss": 0.4464, "step": 16250 }, { "epoch": 1.64, "learning_rate": 4.601760314738223e-05, "loss": 0.4708, "step": 16300 }, { "epoch": 1.65, "learning_rate": 4.6004993442953696e-05, "loss": 0.4666, "step": 16350 }, { "epoch": 1.65, "learning_rate": 4.599238373852517e-05, "loss": 0.4802, "step": 16400 }, { "epoch": 1.66, "learning_rate": 4.5979774034096646e-05, "loss": 0.4958, "step": 16450 }, { "epoch": 1.66, "learning_rate": 4.5967164329668114e-05, "loss": 0.4845, "step": 16500 }, { "epoch": 1.66, "eval_cer": 0.0903212348769295, "eval_loss": 0.534662127494812, "eval_runtime": 1333.9457, "eval_samples_per_second": 6.113, "eval_steps_per_second": 0.765, "step": 16500 }, { "epoch": 1.67, "learning_rate": 4.595455462523959e-05, "loss": 0.4807, "step": 16550 }, { "epoch": 1.67, "learning_rate": 4.594194492081106e-05, "loss": 0.498, "step": 16600 }, { "epoch": 1.68, "learning_rate": 4.592933521638253e-05, "loss": 0.4784, "step": 16650 }, { "epoch": 1.68, "learning_rate": 4.5916725511954e-05, "loss": 0.5014, "step": 16700 }, { "epoch": 1.69, "learning_rate": 4.5904115807525475e-05, "loss": 0.4717, "step": 16750 }, { "epoch": 1.69, "learning_rate": 4.589150610309694e-05, "loss": 0.4813, "step": 16800 }, { "epoch": 1.7, "learning_rate": 4.587889639866842e-05, "loss": 0.4817, "step": 16850 }, { "epoch": 1.7, "learning_rate": 4.586628669423989e-05, "loss": 0.465, "step": 16900 }, { "epoch": 1.71, "learning_rate": 4.585367698981136e-05, "loss": 0.4852, "step": 16950 }, { "epoch": 1.71, "learning_rate": 4.5841067285382836e-05, "loss": 0.4748, "step": 17000 }, { "epoch": 1.71, "eval_cer": 0.0911026843255562, "eval_loss": 0.5354501008987427, "eval_runtime": 1332.928, "eval_samples_per_second": 6.118, "eval_steps_per_second": 0.765, "step": 17000 }, { "epoch": 1.72, "learning_rate": 4.5828457580954304e-05, "loss": 0.4613, "step": 17050 }, { "epoch": 1.72, "learning_rate": 4.581584787652577e-05, "loss": 0.4902, "step": 17100 }, { "epoch": 1.73, "learning_rate": 4.580323817209725e-05, "loss": 0.4604, "step": 17150 }, { "epoch": 1.73, "learning_rate": 4.579062846766872e-05, "loss": 0.4675, "step": 17200 }, { "epoch": 1.74, "learning_rate": 4.5778018763240196e-05, "loss": 0.483, "step": 17250 }, { "epoch": 1.74, "learning_rate": 4.5765409058811664e-05, "loss": 0.4614, "step": 17300 }, { "epoch": 1.75, "learning_rate": 4.575279935438313e-05, "loss": 0.5053, "step": 17350 }, { "epoch": 1.75, "learning_rate": 4.574018964995461e-05, "loss": 0.4748, "step": 17400 }, { "epoch": 1.76, "learning_rate": 4.5727579945526076e-05, "loss": 0.4593, "step": 17450 }, { "epoch": 1.76, "learning_rate": 4.571497024109755e-05, "loss": 0.4683, "step": 17500 }, { "epoch": 1.76, "eval_cer": 0.08885173631218127, "eval_loss": 0.5252342820167542, "eval_runtime": 1414.3948, "eval_samples_per_second": 5.766, "eval_steps_per_second": 0.721, "step": 17500 }, { "epoch": 1.77, "learning_rate": 4.570236053666902e-05, "loss": 0.4376, "step": 17550 }, { "epoch": 1.77, "learning_rate": 4.568975083224049e-05, "loss": 0.4789, "step": 17600 }, { "epoch": 1.78, "learning_rate": 4.567714112781197e-05, "loss": 0.4908, "step": 17650 }, { "epoch": 1.78, "learning_rate": 4.5664531423383436e-05, "loss": 0.4637, "step": 17700 }, { "epoch": 1.79, "learning_rate": 4.565192171895491e-05, "loss": 0.4632, "step": 17750 }, { "epoch": 1.79, "learning_rate": 4.563931201452638e-05, "loss": 0.4688, "step": 17800 }, { "epoch": 1.8, "learning_rate": 4.5626702310097854e-05, "loss": 0.4321, "step": 17850 }, { "epoch": 1.8, "learning_rate": 4.561409260566932e-05, "loss": 0.4646, "step": 17900 }, { "epoch": 1.81, "learning_rate": 4.56014829012408e-05, "loss": 0.4738, "step": 17950 }, { "epoch": 1.81, "learning_rate": 4.558887319681227e-05, "loss": 0.464, "step": 18000 }, { "epoch": 1.81, "eval_cer": 0.08779942589928953, "eval_loss": 0.5221949219703674, "eval_runtime": 1605.2756, "eval_samples_per_second": 5.08, "eval_steps_per_second": 0.635, "step": 18000 }, { "epoch": 1.82, "learning_rate": 4.557626349238374e-05, "loss": 0.4673, "step": 18050 }, { "epoch": 1.82, "learning_rate": 4.5563653787955215e-05, "loss": 0.4961, "step": 18100 }, { "epoch": 1.83, "learning_rate": 4.555104408352668e-05, "loss": 0.4911, "step": 18150 }, { "epoch": 1.83, "learning_rate": 4.553843437909816e-05, "loss": 0.4647, "step": 18200 }, { "epoch": 1.84, "learning_rate": 4.5525824674669626e-05, "loss": 0.4919, "step": 18250 }, { "epoch": 1.84, "learning_rate": 4.5513214970241094e-05, "loss": 0.4738, "step": 18300 }, { "epoch": 1.85, "learning_rate": 4.5500605265812576e-05, "loss": 0.4821, "step": 18350 }, { "epoch": 1.85, "learning_rate": 4.5487995561384044e-05, "loss": 0.4494, "step": 18400 }, { "epoch": 1.86, "learning_rate": 4.547538585695552e-05, "loss": 0.4665, "step": 18450 }, { "epoch": 1.86, "learning_rate": 4.546277615252699e-05, "loss": 0.492, "step": 18500 }, { "epoch": 1.86, "eval_cer": 0.08628322716828872, "eval_loss": 0.5160471200942993, "eval_runtime": 1638.7416, "eval_samples_per_second": 4.976, "eval_steps_per_second": 0.622, "step": 18500 }, { "epoch": 1.87, "learning_rate": 4.5450166448098455e-05, "loss": 0.4561, "step": 18550 }, { "epoch": 1.87, "learning_rate": 4.543755674366993e-05, "loss": 0.4897, "step": 18600 }, { "epoch": 1.88, "learning_rate": 4.54249470392414e-05, "loss": 0.4646, "step": 18650 }, { "epoch": 1.88, "learning_rate": 4.541233733481288e-05, "loss": 0.4708, "step": 18700 }, { "epoch": 1.89, "learning_rate": 4.539972763038435e-05, "loss": 0.4394, "step": 18750 }, { "epoch": 1.89, "learning_rate": 4.5387117925955816e-05, "loss": 0.4455, "step": 18800 }, { "epoch": 1.9, "learning_rate": 4.537450822152729e-05, "loss": 0.415, "step": 18850 }, { "epoch": 1.9, "learning_rate": 4.536189851709876e-05, "loss": 0.4489, "step": 18900 }, { "epoch": 1.91, "learning_rate": 4.5349288812670234e-05, "loss": 0.4546, "step": 18950 }, { "epoch": 1.91, "learning_rate": 4.53366791082417e-05, "loss": 0.473, "step": 19000 }, { "epoch": 1.91, "eval_cer": 0.08715807694942061, "eval_loss": 0.5122098326683044, "eval_runtime": 1509.7705, "eval_samples_per_second": 5.401, "eval_steps_per_second": 0.676, "step": 19000 }, { "epoch": 1.92, "learning_rate": 4.5324069403813177e-05, "loss": 0.4702, "step": 19050 }, { "epoch": 1.92, "learning_rate": 4.531145969938465e-05, "loss": 0.4584, "step": 19100 }, { "epoch": 1.93, "learning_rate": 4.529884999495612e-05, "loss": 0.4431, "step": 19150 }, { "epoch": 1.93, "learning_rate": 4.5286240290527594e-05, "loss": 0.4363, "step": 19200 }, { "epoch": 1.94, "learning_rate": 4.527363058609906e-05, "loss": 0.4495, "step": 19250 }, { "epoch": 1.94, "learning_rate": 4.526102088167054e-05, "loss": 0.4214, "step": 19300 }, { "epoch": 1.95, "learning_rate": 4.5248411177242005e-05, "loss": 0.454, "step": 19350 }, { "epoch": 1.95, "learning_rate": 4.523580147281348e-05, "loss": 0.4675, "step": 19400 }, { "epoch": 1.96, "learning_rate": 4.5223191768384955e-05, "loss": 0.4417, "step": 19450 }, { "epoch": 1.96, "learning_rate": 4.521058206395642e-05, "loss": 0.4664, "step": 19500 }, { "epoch": 1.96, "eval_cer": 0.085579611330083, "eval_loss": 0.5101404190063477, "eval_runtime": 1299.5772, "eval_samples_per_second": 6.275, "eval_steps_per_second": 0.785, "step": 19500 }, { "epoch": 1.97, "learning_rate": 4.51979723595279e-05, "loss": 0.4602, "step": 19550 }, { "epoch": 1.97, "learning_rate": 4.5185362655099366e-05, "loss": 0.4792, "step": 19600 }, { "epoch": 1.98, "learning_rate": 4.517275295067084e-05, "loss": 0.4376, "step": 19650 }, { "epoch": 1.98, "learning_rate": 4.516014324624231e-05, "loss": 0.4495, "step": 19700 }, { "epoch": 1.99, "learning_rate": 4.514753354181378e-05, "loss": 0.4408, "step": 19750 }, { "epoch": 1.99, "learning_rate": 4.513492383738525e-05, "loss": 0.4566, "step": 19800 }, { "epoch": 2.0, "learning_rate": 4.512231413295673e-05, "loss": 0.4536, "step": 19850 }, { "epoch": 2.0, "learning_rate": 4.51097044285282e-05, "loss": 0.3997, "step": 19900 }, { "epoch": 2.01, "learning_rate": 4.509709472409967e-05, "loss": 0.3247, "step": 19950 }, { "epoch": 2.01, "learning_rate": 4.508473721375971e-05, "loss": 0.3219, "step": 20000 }, { "epoch": 2.01, "eval_cer": 0.08687164926307138, "eval_loss": 0.5130881071090698, "eval_runtime": 1325.7684, "eval_samples_per_second": 6.151, "eval_steps_per_second": 0.769, "step": 20000 }, { "epoch": 2.02, "learning_rate": 4.507212750933118e-05, "loss": 0.3348, "step": 20050 }, { "epoch": 2.02, "learning_rate": 4.505951780490266e-05, "loss": 0.3511, "step": 20100 }, { "epoch": 2.03, "learning_rate": 4.5046908100474126e-05, "loss": 0.3458, "step": 20150 }, { "epoch": 2.03, "learning_rate": 4.50342983960456e-05, "loss": 0.3256, "step": 20200 }, { "epoch": 2.04, "learning_rate": 4.502168869161707e-05, "loss": 0.3145, "step": 20250 }, { "epoch": 2.04, "learning_rate": 4.5009078987188544e-05, "loss": 0.3289, "step": 20300 }, { "epoch": 2.05, "learning_rate": 4.499646928276001e-05, "loss": 0.3326, "step": 20350 }, { "epoch": 2.05, "learning_rate": 4.4983859578331487e-05, "loss": 0.3355, "step": 20400 }, { "epoch": 2.06, "learning_rate": 4.497124987390296e-05, "loss": 0.3591, "step": 20450 }, { "epoch": 2.06, "learning_rate": 4.495864016947443e-05, "loss": 0.3183, "step": 20500 }, { "epoch": 2.06, "eval_cer": 0.08809208027447245, "eval_loss": 0.507744550704956, "eval_runtime": 1333.9002, "eval_samples_per_second": 6.114, "eval_steps_per_second": 0.765, "step": 20500 }, { "epoch": 2.07, "learning_rate": 4.4946030465045904e-05, "loss": 0.3284, "step": 20550 }, { "epoch": 2.07, "learning_rate": 4.493342076061737e-05, "loss": 0.3228, "step": 20600 }, { "epoch": 2.08, "learning_rate": 4.492081105618884e-05, "loss": 0.3428, "step": 20650 }, { "epoch": 2.08, "learning_rate": 4.4908201351760316e-05, "loss": 0.3438, "step": 20700 }, { "epoch": 2.09, "learning_rate": 4.4895591647331784e-05, "loss": 0.3442, "step": 20750 }, { "epoch": 2.09, "learning_rate": 4.4882981942903265e-05, "loss": 0.3218, "step": 20800 }, { "epoch": 2.1, "learning_rate": 4.487037223847473e-05, "loss": 0.3503, "step": 20850 }, { "epoch": 2.1, "learning_rate": 4.485776253404621e-05, "loss": 0.3304, "step": 20900 }, { "epoch": 2.11, "learning_rate": 4.4845152829617676e-05, "loss": 0.3257, "step": 20950 }, { "epoch": 2.11, "learning_rate": 4.4832543125189144e-05, "loss": 0.3508, "step": 21000 }, { "epoch": 2.11, "eval_cer": 0.0858722657052659, "eval_loss": 0.5087429881095886, "eval_runtime": 1663.7878, "eval_samples_per_second": 4.901, "eval_steps_per_second": 0.613, "step": 21000 }, { "epoch": 2.12, "learning_rate": 4.481993342076062e-05, "loss": 0.3515, "step": 21050 }, { "epoch": 2.12, "learning_rate": 4.480732371633209e-05, "loss": 0.3354, "step": 21100 }, { "epoch": 2.13, "learning_rate": 4.479471401190357e-05, "loss": 0.3533, "step": 21150 }, { "epoch": 2.13, "learning_rate": 4.478210430747504e-05, "loss": 0.3772, "step": 21200 }, { "epoch": 2.14, "learning_rate": 4.4769494603046505e-05, "loss": 0.3351, "step": 21250 }, { "epoch": 2.14, "learning_rate": 4.475688489861798e-05, "loss": 0.3451, "step": 21300 }, { "epoch": 2.15, "learning_rate": 4.474427519418945e-05, "loss": 0.3494, "step": 21350 }, { "epoch": 2.15, "learning_rate": 4.473166548976092e-05, "loss": 0.3366, "step": 21400 }, { "epoch": 2.16, "learning_rate": 4.471905578533239e-05, "loss": 0.3487, "step": 21450 }, { "epoch": 2.16, "learning_rate": 4.4706446080903866e-05, "loss": 0.3671, "step": 21500 }, { "epoch": 2.16, "eval_cer": 0.08502232267946874, "eval_loss": 0.507188081741333, "eval_runtime": 1648.1593, "eval_samples_per_second": 4.948, "eval_steps_per_second": 0.619, "step": 21500 }, { "epoch": 2.17, "learning_rate": 4.469383637647534e-05, "loss": 0.3449, "step": 21550 }, { "epoch": 2.17, "learning_rate": 4.468122667204681e-05, "loss": 0.3232, "step": 21600 }, { "epoch": 2.18, "learning_rate": 4.4668616967618284e-05, "loss": 0.3534, "step": 21650 }, { "epoch": 2.18, "learning_rate": 4.465600726318975e-05, "loss": 0.3388, "step": 21700 }, { "epoch": 2.19, "learning_rate": 4.464339755876123e-05, "loss": 0.3602, "step": 21750 }, { "epoch": 2.19, "learning_rate": 4.4630787854332695e-05, "loss": 0.3342, "step": 21800 }, { "epoch": 2.2, "learning_rate": 4.461817814990416e-05, "loss": 0.3483, "step": 21850 }, { "epoch": 2.2, "learning_rate": 4.4605568445475645e-05, "loss": 0.3479, "step": 21900 }, { "epoch": 2.21, "learning_rate": 4.459295874104711e-05, "loss": 0.3509, "step": 21950 }, { "epoch": 2.21, "learning_rate": 4.458034903661859e-05, "loss": 0.3512, "step": 22000 }, { "epoch": 2.21, "eval_cer": 0.0838672719008213, "eval_loss": 0.506687581539154, "eval_runtime": 1638.6031, "eval_samples_per_second": 4.977, "eval_steps_per_second": 0.622, "step": 22000 }, { "epoch": 2.22, "learning_rate": 4.4567739332190056e-05, "loss": 0.345, "step": 22050 }, { "epoch": 2.22, "learning_rate": 4.4555129627761524e-05, "loss": 0.3282, "step": 22100 }, { "epoch": 2.23, "learning_rate": 4.4542519923333e-05, "loss": 0.3641, "step": 22150 }, { "epoch": 2.23, "learning_rate": 4.452991021890447e-05, "loss": 0.3421, "step": 22200 }, { "epoch": 2.24, "learning_rate": 4.451730051447594e-05, "loss": 0.3437, "step": 22250 }, { "epoch": 2.24, "learning_rate": 4.4504690810047417e-05, "loss": 0.3428, "step": 22300 }, { "epoch": 2.25, "learning_rate": 4.4492081105618885e-05, "loss": 0.3419, "step": 22350 }, { "epoch": 2.25, "learning_rate": 4.447947140119036e-05, "loss": 0.3525, "step": 22400 }, { "epoch": 2.26, "learning_rate": 4.446686169676183e-05, "loss": 0.3528, "step": 22450 }, { "epoch": 2.26, "learning_rate": 4.44542519923333e-05, "loss": 0.3399, "step": 22500 }, { "epoch": 2.26, "eval_cer": 0.08266240761150442, "eval_loss": 0.4957926571369171, "eval_runtime": 1315.8531, "eval_samples_per_second": 6.198, "eval_steps_per_second": 0.775, "step": 22500 }, { "epoch": 2.27, "learning_rate": 4.444164228790477e-05, "loss": 0.3288, "step": 22550 }, { "epoch": 2.27, "learning_rate": 4.4429032583476245e-05, "loss": 0.3497, "step": 22600 }, { "epoch": 2.28, "learning_rate": 4.441642287904772e-05, "loss": 0.3303, "step": 22650 }, { "epoch": 2.28, "learning_rate": 4.440381317461919e-05, "loss": 0.3446, "step": 22700 }, { "epoch": 2.29, "learning_rate": 4.439120347019066e-05, "loss": 0.3517, "step": 22750 }, { "epoch": 2.29, "learning_rate": 4.437859376576213e-05, "loss": 0.3723, "step": 22800 }, { "epoch": 2.3, "learning_rate": 4.4365984061333606e-05, "loss": 0.3354, "step": 22850 }, { "epoch": 2.3, "learning_rate": 4.4353374356905074e-05, "loss": 0.3323, "step": 22900 }, { "epoch": 2.31, "learning_rate": 4.434076465247655e-05, "loss": 0.337, "step": 22950 }, { "epoch": 2.31, "learning_rate": 4.432815494804802e-05, "loss": 0.358, "step": 23000 }, { "epoch": 2.31, "eval_cer": 0.0816256639206969, "eval_loss": 0.4962875545024872, "eval_runtime": 1639.4963, "eval_samples_per_second": 4.974, "eval_steps_per_second": 0.622, "step": 23000 }, { "epoch": 2.32, "learning_rate": 4.431554524361949e-05, "loss": 0.3314, "step": 23050 }, { "epoch": 2.32, "learning_rate": 4.430293553919097e-05, "loss": 0.3191, "step": 23100 }, { "epoch": 2.33, "learning_rate": 4.4290578028851005e-05, "loss": 0.3511, "step": 23150 }, { "epoch": 2.33, "learning_rate": 4.427796832442247e-05, "loss": 0.3559, "step": 23200 }, { "epoch": 2.34, "learning_rate": 4.4265358619993955e-05, "loss": 0.3516, "step": 23250 }, { "epoch": 2.34, "learning_rate": 4.425274891556542e-05, "loss": 0.3303, "step": 23300 }, { "epoch": 2.35, "learning_rate": 4.424013921113689e-05, "loss": 0.356, "step": 23350 }, { "epoch": 2.35, "learning_rate": 4.4227529506708366e-05, "loss": 0.3615, "step": 23400 }, { "epoch": 2.36, "learning_rate": 4.4214919802279834e-05, "loss": 0.3393, "step": 23450 }, { "epoch": 2.36, "learning_rate": 4.420231009785131e-05, "loss": 0.3565, "step": 23500 }, { "epoch": 2.36, "eval_cer": 0.08242579343582464, "eval_loss": 0.497714102268219, "eval_runtime": 1368.4674, "eval_samples_per_second": 5.959, "eval_steps_per_second": 0.745, "step": 23500 }, { "epoch": 2.37, "learning_rate": 4.418970039342278e-05, "loss": 0.3758, "step": 23550 }, { "epoch": 2.37, "learning_rate": 4.417709068899425e-05, "loss": 0.3219, "step": 23600 }, { "epoch": 2.38, "learning_rate": 4.4164480984565727e-05, "loss": 0.3499, "step": 23650 }, { "epoch": 2.38, "learning_rate": 4.4151871280137195e-05, "loss": 0.3497, "step": 23700 }, { "epoch": 2.39, "learning_rate": 4.413926157570867e-05, "loss": 0.3743, "step": 23750 }, { "epoch": 2.39, "learning_rate": 4.412665187128014e-05, "loss": 0.3465, "step": 23800 }, { "epoch": 2.4, "learning_rate": 4.411404216685161e-05, "loss": 0.3478, "step": 23850 }, { "epoch": 2.4, "learning_rate": 4.410143246242308e-05, "loss": 0.3502, "step": 23900 }, { "epoch": 2.41, "learning_rate": 4.4088822757994555e-05, "loss": 0.3285, "step": 23950 }, { "epoch": 2.41, "learning_rate": 4.407621305356603e-05, "loss": 0.3419, "step": 24000 }, { "epoch": 2.41, "eval_cer": 0.08279628142142853, "eval_loss": 0.4959103465080261, "eval_runtime": 1557.8526, "eval_samples_per_second": 5.235, "eval_steps_per_second": 0.655, "step": 24000 }, { "epoch": 2.42, "learning_rate": 4.40636033491375e-05, "loss": 0.3607, "step": 24050 }, { "epoch": 2.43, "learning_rate": 4.405099364470897e-05, "loss": 0.3641, "step": 24100 }, { "epoch": 2.43, "learning_rate": 4.403838394028044e-05, "loss": 0.3601, "step": 24150 }, { "epoch": 2.44, "learning_rate": 4.4025774235851916e-05, "loss": 0.3435, "step": 24200 }, { "epoch": 2.44, "learning_rate": 4.4013164531423384e-05, "loss": 0.3455, "step": 24250 }, { "epoch": 2.45, "learning_rate": 4.400055482699485e-05, "loss": 0.3498, "step": 24300 }, { "epoch": 2.45, "learning_rate": 4.398794512256633e-05, "loss": 0.3493, "step": 24350 }, { "epoch": 2.46, "learning_rate": 4.39753354181378e-05, "loss": 0.3246, "step": 24400 }, { "epoch": 2.46, "learning_rate": 4.396272571370928e-05, "loss": 0.3162, "step": 24450 }, { "epoch": 2.47, "learning_rate": 4.3950116009280745e-05, "loss": 0.3856, "step": 24500 }, { "epoch": 2.47, "eval_cer": 0.08169104415345052, "eval_loss": 0.4946542978286743, "eval_runtime": 1342.4827, "eval_samples_per_second": 6.075, "eval_steps_per_second": 0.76, "step": 24500 }, { "epoch": 2.47, "learning_rate": 4.393750630485221e-05, "loss": 0.357, "step": 24550 }, { "epoch": 2.48, "learning_rate": 4.392489660042369e-05, "loss": 0.3553, "step": 24600 }, { "epoch": 2.48, "learning_rate": 4.3912286895995156e-05, "loss": 0.3516, "step": 24650 }, { "epoch": 2.49, "learning_rate": 4.389967719156663e-05, "loss": 0.3251, "step": 24700 }, { "epoch": 2.49, "learning_rate": 4.3887067487138106e-05, "loss": 0.3543, "step": 24750 }, { "epoch": 2.5, "learning_rate": 4.3874457782709574e-05, "loss": 0.3342, "step": 24800 }, { "epoch": 2.5, "learning_rate": 4.386184807828105e-05, "loss": 0.3502, "step": 24850 }, { "epoch": 2.51, "learning_rate": 4.384923837385252e-05, "loss": 0.3539, "step": 24900 }, { "epoch": 2.51, "learning_rate": 4.383662866942399e-05, "loss": 0.3539, "step": 24950 }, { "epoch": 2.52, "learning_rate": 4.382401896499546e-05, "loss": 0.3601, "step": 25000 }, { "epoch": 2.52, "eval_cer": 0.08011880522294659, "eval_loss": 0.48510172963142395, "eval_runtime": 1322.8436, "eval_samples_per_second": 6.165, "eval_steps_per_second": 0.771, "step": 25000 }, { "epoch": 2.52, "learning_rate": 4.3811409260566935e-05, "loss": 0.325, "step": 25050 }, { "epoch": 2.53, "learning_rate": 4.37987995561384e-05, "loss": 0.334, "step": 25100 }, { "epoch": 2.53, "learning_rate": 4.378618985170988e-05, "loss": 0.3448, "step": 25150 }, { "epoch": 2.54, "learning_rate": 4.377383234136992e-05, "loss": 0.3851, "step": 25200 }, { "epoch": 2.54, "learning_rate": 4.376122263694139e-05, "loss": 0.3414, "step": 25250 }, { "epoch": 2.55, "learning_rate": 4.374861293251286e-05, "loss": 0.3713, "step": 25300 }, { "epoch": 2.55, "learning_rate": 4.373600322808434e-05, "loss": 0.3527, "step": 25350 }, { "epoch": 2.56, "learning_rate": 4.372339352365581e-05, "loss": 0.3463, "step": 25400 }, { "epoch": 2.56, "learning_rate": 4.371078381922728e-05, "loss": 0.3376, "step": 25450 }, { "epoch": 2.57, "learning_rate": 4.369817411479875e-05, "loss": 0.3427, "step": 25500 }, { "epoch": 2.57, "eval_cer": 0.08287100168743268, "eval_loss": 0.4828871488571167, "eval_runtime": 1335.9004, "eval_samples_per_second": 6.104, "eval_steps_per_second": 0.764, "step": 25500 }, { "epoch": 2.57, "learning_rate": 4.368556441037022e-05, "loss": 0.3311, "step": 25550 }, { "epoch": 2.58, "learning_rate": 4.3672954705941694e-05, "loss": 0.3592, "step": 25600 }, { "epoch": 2.58, "learning_rate": 4.366034500151316e-05, "loss": 0.3442, "step": 25650 }, { "epoch": 2.59, "learning_rate": 4.3647735297084644e-05, "loss": 0.3381, "step": 25700 }, { "epoch": 2.59, "learning_rate": 4.363512559265611e-05, "loss": 0.3632, "step": 25750 }, { "epoch": 2.6, "learning_rate": 4.362251588822758e-05, "loss": 0.3482, "step": 25800 }, { "epoch": 2.6, "learning_rate": 4.3609906183799055e-05, "loss": 0.3627, "step": 25850 }, { "epoch": 2.61, "learning_rate": 4.359729647937052e-05, "loss": 0.3439, "step": 25900 }, { "epoch": 2.61, "learning_rate": 4.3584686774942e-05, "loss": 0.3522, "step": 25950 }, { "epoch": 2.62, "learning_rate": 4.3572077070513466e-05, "loss": 0.3676, "step": 26000 }, { "epoch": 2.62, "eval_cer": 0.0803958928760453, "eval_loss": 0.48255375027656555, "eval_runtime": 1330.4175, "eval_samples_per_second": 6.13, "eval_steps_per_second": 0.767, "step": 26000 }, { "epoch": 2.62, "learning_rate": 4.355946736608494e-05, "loss": 0.3437, "step": 26050 }, { "epoch": 2.63, "learning_rate": 4.3546857661656416e-05, "loss": 0.3466, "step": 26100 }, { "epoch": 2.63, "learning_rate": 4.3534247957227884e-05, "loss": 0.3096, "step": 26150 }, { "epoch": 2.64, "learning_rate": 4.352163825279936e-05, "loss": 0.3484, "step": 26200 }, { "epoch": 2.64, "learning_rate": 4.350902854837083e-05, "loss": 0.3596, "step": 26250 }, { "epoch": 2.65, "learning_rate": 4.34964188439423e-05, "loss": 0.3363, "step": 26300 }, { "epoch": 2.65, "learning_rate": 4.348380913951377e-05, "loss": 0.3616, "step": 26350 }, { "epoch": 2.66, "learning_rate": 4.347119943508524e-05, "loss": 0.3668, "step": 26400 }, { "epoch": 2.66, "learning_rate": 4.345858973065672e-05, "loss": 0.3552, "step": 26450 }, { "epoch": 2.67, "learning_rate": 4.344598002622819e-05, "loss": 0.3691, "step": 26500 }, { "epoch": 2.67, "eval_cer": 0.07944320948449243, "eval_loss": 0.47338464856147766, "eval_runtime": 1393.5786, "eval_samples_per_second": 5.852, "eval_steps_per_second": 0.732, "step": 26500 }, { "epoch": 2.67, "learning_rate": 4.343337032179966e-05, "loss": 0.3451, "step": 26550 }, { "epoch": 2.68, "learning_rate": 4.342076061737113e-05, "loss": 0.3272, "step": 26600 }, { "epoch": 2.68, "learning_rate": 4.34081509129426e-05, "loss": 0.3329, "step": 26650 }, { "epoch": 2.69, "learning_rate": 4.3395541208514074e-05, "loss": 0.3281, "step": 26700 }, { "epoch": 2.69, "learning_rate": 4.338293150408554e-05, "loss": 0.3423, "step": 26750 }, { "epoch": 2.7, "learning_rate": 4.337032179965702e-05, "loss": 0.3522, "step": 26800 }, { "epoch": 2.7, "learning_rate": 4.335771209522849e-05, "loss": 0.3457, "step": 26850 }, { "epoch": 2.71, "learning_rate": 4.3345102390799967e-05, "loss": 0.3587, "step": 26900 }, { "epoch": 2.71, "learning_rate": 4.3332492686371435e-05, "loss": 0.3545, "step": 26950 }, { "epoch": 2.72, "learning_rate": 4.33198829819429e-05, "loss": 0.3541, "step": 27000 }, { "epoch": 2.72, "eval_cer": 0.08026513241053805, "eval_loss": 0.4757711887359619, "eval_runtime": 1337.5155, "eval_samples_per_second": 6.097, "eval_steps_per_second": 0.763, "step": 27000 }, { "epoch": 2.72, "learning_rate": 4.330727327751438e-05, "loss": 0.3502, "step": 27050 }, { "epoch": 2.73, "learning_rate": 4.3294663573085846e-05, "loss": 0.34, "step": 27100 }, { "epoch": 2.73, "learning_rate": 4.328205386865732e-05, "loss": 0.3455, "step": 27150 }, { "epoch": 2.74, "learning_rate": 4.3269444164228795e-05, "loss": 0.376, "step": 27200 }, { "epoch": 2.74, "learning_rate": 4.325708665388883e-05, "loss": 0.3353, "step": 27250 }, { "epoch": 2.75, "learning_rate": 4.324447694946031e-05, "loss": 0.3172, "step": 27300 }, { "epoch": 2.75, "learning_rate": 4.3231867245031776e-05, "loss": 0.3513, "step": 27350 }, { "epoch": 2.76, "learning_rate": 4.321925754060325e-05, "loss": 0.3554, "step": 27400 }, { "epoch": 2.76, "learning_rate": 4.3206647836174726e-05, "loss": 0.3535, "step": 27450 }, { "epoch": 2.77, "learning_rate": 4.3194038131746194e-05, "loss": 0.3593, "step": 27500 }, { "epoch": 2.77, "eval_cer": 0.08003785826810877, "eval_loss": 0.4731917083263397, "eval_runtime": 1382.8241, "eval_samples_per_second": 5.897, "eval_steps_per_second": 0.738, "step": 27500 }, { "epoch": 2.77, "learning_rate": 4.318142842731767e-05, "loss": 0.3155, "step": 27550 }, { "epoch": 2.78, "learning_rate": 4.316881872288914e-05, "loss": 0.3483, "step": 27600 }, { "epoch": 2.78, "learning_rate": 4.3156209018460605e-05, "loss": 0.3657, "step": 27650 }, { "epoch": 2.79, "learning_rate": 4.314359931403208e-05, "loss": 0.3381, "step": 27700 }, { "epoch": 2.79, "learning_rate": 4.313098960960355e-05, "loss": 0.3528, "step": 27750 }, { "epoch": 2.8, "learning_rate": 4.311837990517503e-05, "loss": 0.3465, "step": 27800 }, { "epoch": 2.8, "learning_rate": 4.31057702007465e-05, "loss": 0.3259, "step": 27850 }, { "epoch": 2.81, "learning_rate": 4.3093160496317966e-05, "loss": 0.344, "step": 27900 }, { "epoch": 2.81, "learning_rate": 4.308055079188944e-05, "loss": 0.3261, "step": 27950 }, { "epoch": 2.82, "learning_rate": 4.306794108746091e-05, "loss": 0.3482, "step": 28000 }, { "epoch": 2.82, "eval_cer": 0.08137036967851606, "eval_loss": 0.4740140736103058, "eval_runtime": 1600.2061, "eval_samples_per_second": 5.096, "eval_steps_per_second": 0.637, "step": 28000 }, { "epoch": 2.82, "learning_rate": 4.3055331383032384e-05, "loss": 0.3398, "step": 28050 }, { "epoch": 2.83, "learning_rate": 4.304272167860385e-05, "loss": 0.3321, "step": 28100 }, { "epoch": 2.83, "learning_rate": 4.303011197417533e-05, "loss": 0.3357, "step": 28150 }, { "epoch": 2.84, "learning_rate": 4.30175022697468e-05, "loss": 0.3682, "step": 28200 }, { "epoch": 2.84, "learning_rate": 4.300489256531827e-05, "loss": 0.3796, "step": 28250 }, { "epoch": 2.85, "learning_rate": 4.2992282860889745e-05, "loss": 0.3592, "step": 28300 }, { "epoch": 2.85, "learning_rate": 4.297967315646121e-05, "loss": 0.3589, "step": 28350 }, { "epoch": 2.86, "learning_rate": 4.296706345203269e-05, "loss": 0.3341, "step": 28400 }, { "epoch": 2.86, "learning_rate": 4.2954453747604156e-05, "loss": 0.361, "step": 28450 }, { "epoch": 2.87, "learning_rate": 4.294184404317563e-05, "loss": 0.3382, "step": 28500 }, { "epoch": 2.87, "eval_cer": 0.07849363943735639, "eval_loss": 0.465993195772171, "eval_runtime": 1313.6382, "eval_samples_per_second": 6.208, "eval_steps_per_second": 0.776, "step": 28500 }, { "epoch": 2.87, "learning_rate": 4.2929234338747105e-05, "loss": 0.349, "step": 28550 }, { "epoch": 2.88, "learning_rate": 4.2916624634318574e-05, "loss": 0.3667, "step": 28600 }, { "epoch": 2.88, "learning_rate": 4.290401492989005e-05, "loss": 0.3702, "step": 28650 }, { "epoch": 2.89, "learning_rate": 4.2891405225461517e-05, "loss": 0.3469, "step": 28700 }, { "epoch": 2.89, "learning_rate": 4.287879552103299e-05, "loss": 0.3605, "step": 28750 }, { "epoch": 2.9, "learning_rate": 4.286618581660446e-05, "loss": 0.3427, "step": 28800 }, { "epoch": 2.9, "learning_rate": 4.285357611217593e-05, "loss": 0.3641, "step": 28850 }, { "epoch": 2.91, "learning_rate": 4.284096640774741e-05, "loss": 0.3348, "step": 28900 }, { "epoch": 2.91, "learning_rate": 4.282835670331888e-05, "loss": 0.3494, "step": 28950 }, { "epoch": 2.92, "learning_rate": 4.281574699889035e-05, "loss": 0.359, "step": 29000 }, { "epoch": 2.92, "eval_cer": 0.07920970865322947, "eval_loss": 0.4593048691749573, "eval_runtime": 1321.829, "eval_samples_per_second": 6.169, "eval_steps_per_second": 0.772, "step": 29000 }, { "epoch": 2.92, "learning_rate": 4.280313729446182e-05, "loss": 0.3485, "step": 29050 }, { "epoch": 2.93, "learning_rate": 4.279052759003329e-05, "loss": 0.3576, "step": 29100 }, { "epoch": 2.93, "learning_rate": 4.277791788560476e-05, "loss": 0.3261, "step": 29150 }, { "epoch": 2.94, "learning_rate": 4.276530818117623e-05, "loss": 0.348, "step": 29200 }, { "epoch": 2.94, "learning_rate": 4.2752698476747706e-05, "loss": 0.3582, "step": 29250 }, { "epoch": 2.95, "learning_rate": 4.274008877231918e-05, "loss": 0.3486, "step": 29300 }, { "epoch": 2.95, "learning_rate": 4.272747906789065e-05, "loss": 0.3439, "step": 29350 }, { "epoch": 2.96, "learning_rate": 4.2714869363462124e-05, "loss": 0.3515, "step": 29400 }, { "epoch": 2.96, "learning_rate": 4.270225965903359e-05, "loss": 0.3398, "step": 29450 }, { "epoch": 2.97, "learning_rate": 4.268990214869364e-05, "loss": 0.3382, "step": 29500 }, { "epoch": 2.97, "eval_cer": 0.07829749873909551, "eval_loss": 0.4634835124015808, "eval_runtime": 1315.1087, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 29500 }, { "epoch": 2.97, "learning_rate": 4.267729244426511e-05, "loss": 0.3532, "step": 29550 }, { "epoch": 2.98, "learning_rate": 4.266468273983658e-05, "loss": 0.3239, "step": 29600 }, { "epoch": 2.98, "learning_rate": 4.2652073035408055e-05, "loss": 0.3273, "step": 29650 }, { "epoch": 2.99, "learning_rate": 4.263946333097952e-05, "loss": 0.3458, "step": 29700 }, { "epoch": 2.99, "learning_rate": 4.2626853626551e-05, "loss": 0.3396, "step": 29750 }, { "epoch": 3.0, "learning_rate": 4.2614243922122466e-05, "loss": 0.3191, "step": 29800 }, { "epoch": 3.0, "learning_rate": 4.260163421769394e-05, "loss": 0.2726, "step": 29850 }, { "epoch": 3.01, "learning_rate": 4.2589024513265416e-05, "loss": 0.231, "step": 29900 }, { "epoch": 3.01, "learning_rate": 4.2576414808836884e-05, "loss": 0.2419, "step": 29950 }, { "epoch": 3.02, "learning_rate": 4.256380510440836e-05, "loss": 0.2164, "step": 30000 }, { "epoch": 3.02, "eval_cer": 0.07885790073412662, "eval_loss": 0.47512373328208923, "eval_runtime": 1318.9572, "eval_samples_per_second": 6.183, "eval_steps_per_second": 0.773, "step": 30000 }, { "epoch": 3.02, "learning_rate": 4.2551195399979827e-05, "loss": 0.2179, "step": 30050 }, { "epoch": 3.03, "learning_rate": 4.2538585695551295e-05, "loss": 0.2439, "step": 30100 }, { "epoch": 3.03, "learning_rate": 4.252597599112277e-05, "loss": 0.2332, "step": 30150 }, { "epoch": 3.04, "learning_rate": 4.251336628669424e-05, "loss": 0.2113, "step": 30200 }, { "epoch": 3.04, "learning_rate": 4.250075658226572e-05, "loss": 0.2481, "step": 30250 }, { "epoch": 3.05, "learning_rate": 4.248814687783719e-05, "loss": 0.2528, "step": 30300 }, { "epoch": 3.05, "learning_rate": 4.2475537173408655e-05, "loss": 0.2551, "step": 30350 }, { "epoch": 3.06, "learning_rate": 4.246292746898013e-05, "loss": 0.2384, "step": 30400 }, { "epoch": 3.06, "learning_rate": 4.24503177645516e-05, "loss": 0.2488, "step": 30450 }, { "epoch": 3.07, "learning_rate": 4.243770806012307e-05, "loss": 0.2449, "step": 30500 }, { "epoch": 3.07, "eval_cer": 0.07689026706268408, "eval_loss": 0.47066158056259155, "eval_runtime": 1305.7328, "eval_samples_per_second": 6.246, "eval_steps_per_second": 0.781, "step": 30500 }, { "epoch": 3.07, "learning_rate": 4.242509835569454e-05, "loss": 0.2353, "step": 30550 }, { "epoch": 3.08, "learning_rate": 4.2412488651266016e-05, "loss": 0.2323, "step": 30600 }, { "epoch": 3.08, "learning_rate": 4.239987894683749e-05, "loss": 0.2416, "step": 30650 }, { "epoch": 3.09, "learning_rate": 4.238726924240896e-05, "loss": 0.2409, "step": 30700 }, { "epoch": 3.09, "learning_rate": 4.2374659537980434e-05, "loss": 0.233, "step": 30750 }, { "epoch": 3.1, "learning_rate": 4.23620498335519e-05, "loss": 0.2239, "step": 30800 }, { "epoch": 3.1, "learning_rate": 4.234944012912338e-05, "loss": 0.2278, "step": 30850 }, { "epoch": 3.11, "learning_rate": 4.2336830424694845e-05, "loss": 0.2502, "step": 30900 }, { "epoch": 3.11, "learning_rate": 4.232422072026631e-05, "loss": 0.2405, "step": 30950 }, { "epoch": 3.12, "learning_rate": 4.2311611015837795e-05, "loss": 0.238, "step": 31000 }, { "epoch": 3.12, "eval_cer": 0.07886101407854346, "eval_loss": 0.4778068959712982, "eval_runtime": 1362.8163, "eval_samples_per_second": 5.984, "eval_steps_per_second": 0.748, "step": 31000 }, { "epoch": 3.12, "learning_rate": 4.229900131140926e-05, "loss": 0.2386, "step": 31050 }, { "epoch": 3.13, "learning_rate": 4.228639160698074e-05, "loss": 0.2668, "step": 31100 }, { "epoch": 3.13, "learning_rate": 4.2273781902552206e-05, "loss": 0.2435, "step": 31150 }, { "epoch": 3.14, "learning_rate": 4.2261172198123674e-05, "loss": 0.2511, "step": 31200 }, { "epoch": 3.14, "learning_rate": 4.224856249369515e-05, "loss": 0.239, "step": 31250 }, { "epoch": 3.15, "learning_rate": 4.223595278926662e-05, "loss": 0.2676, "step": 31300 }, { "epoch": 3.15, "learning_rate": 4.222334308483809e-05, "loss": 0.2245, "step": 31350 }, { "epoch": 3.16, "learning_rate": 4.221073338040957e-05, "loss": 0.2543, "step": 31400 }, { "epoch": 3.16, "learning_rate": 4.219812367598104e-05, "loss": 0.2386, "step": 31450 }, { "epoch": 3.17, "learning_rate": 4.218551397155251e-05, "loss": 0.2364, "step": 31500 }, { "epoch": 3.17, "eval_cer": 0.07789899065374006, "eval_loss": 0.48225945234298706, "eval_runtime": 1331.2304, "eval_samples_per_second": 6.126, "eval_steps_per_second": 0.766, "step": 31500 }, { "epoch": 3.17, "learning_rate": 4.217290426712398e-05, "loss": 0.2336, "step": 31550 }, { "epoch": 3.18, "learning_rate": 4.216029456269545e-05, "loss": 0.2538, "step": 31600 }, { "epoch": 3.18, "learning_rate": 4.214768485826692e-05, "loss": 0.2356, "step": 31650 }, { "epoch": 3.19, "learning_rate": 4.2135075153838396e-05, "loss": 0.2435, "step": 31700 }, { "epoch": 3.19, "learning_rate": 4.212246544940987e-05, "loss": 0.2618, "step": 31750 }, { "epoch": 3.2, "learning_rate": 4.210985574498134e-05, "loss": 0.2521, "step": 31800 }, { "epoch": 3.2, "learning_rate": 4.2097246040552814e-05, "loss": 0.2549, "step": 31850 }, { "epoch": 3.21, "learning_rate": 4.208463633612428e-05, "loss": 0.254, "step": 31900 }, { "epoch": 3.21, "learning_rate": 4.2072026631695757e-05, "loss": 0.2558, "step": 31950 }, { "epoch": 3.22, "learning_rate": 4.2059416927267225e-05, "loss": 0.2633, "step": 32000 }, { "epoch": 3.22, "eval_cer": 0.0768373402075978, "eval_loss": 0.47122320532798767, "eval_runtime": 1302.9545, "eval_samples_per_second": 6.259, "eval_steps_per_second": 0.783, "step": 32000 }, { "epoch": 3.22, "learning_rate": 4.20468072228387e-05, "loss": 0.2562, "step": 32050 }, { "epoch": 3.23, "learning_rate": 4.203419751841017e-05, "loss": 0.2367, "step": 32100 }, { "epoch": 3.24, "learning_rate": 4.202158781398164e-05, "loss": 0.2595, "step": 32150 }, { "epoch": 3.24, "learning_rate": 4.200897810955312e-05, "loss": 0.2439, "step": 32200 }, { "epoch": 3.25, "learning_rate": 4.1996368405124585e-05, "loss": 0.2461, "step": 32250 }, { "epoch": 3.25, "learning_rate": 4.198375870069606e-05, "loss": 0.2591, "step": 32300 }, { "epoch": 3.26, "learning_rate": 4.197114899626753e-05, "loss": 0.2529, "step": 32350 }, { "epoch": 3.26, "learning_rate": 4.1958539291838996e-05, "loss": 0.2688, "step": 32400 }, { "epoch": 3.27, "learning_rate": 4.194592958741047e-05, "loss": 0.2457, "step": 32450 }, { "epoch": 3.27, "learning_rate": 4.1933319882981946e-05, "loss": 0.2483, "step": 32500 }, { "epoch": 3.27, "eval_cer": 0.07903224802146962, "eval_loss": 0.47884607315063477, "eval_runtime": 1326.8262, "eval_samples_per_second": 6.146, "eval_steps_per_second": 0.769, "step": 32500 }, { "epoch": 3.28, "learning_rate": 4.192071017855342e-05, "loss": 0.2431, "step": 32550 }, { "epoch": 3.28, "learning_rate": 4.190810047412489e-05, "loss": 0.2466, "step": 32600 }, { "epoch": 3.29, "learning_rate": 4.189549076969636e-05, "loss": 0.246, "step": 32650 }, { "epoch": 3.29, "learning_rate": 4.188288106526783e-05, "loss": 0.2441, "step": 32700 }, { "epoch": 3.3, "learning_rate": 4.187052355492788e-05, "loss": 0.2392, "step": 32750 }, { "epoch": 3.3, "learning_rate": 4.1857913850499345e-05, "loss": 0.2535, "step": 32800 }, { "epoch": 3.31, "learning_rate": 4.184530414607082e-05, "loss": 0.2715, "step": 32850 }, { "epoch": 3.31, "learning_rate": 4.183269444164229e-05, "loss": 0.2576, "step": 32900 }, { "epoch": 3.32, "learning_rate": 4.182008473721376e-05, "loss": 0.2604, "step": 32950 }, { "epoch": 3.32, "learning_rate": 4.180747503278523e-05, "loss": 0.2429, "step": 33000 }, { "epoch": 3.32, "eval_cer": 0.07802663777483047, "eval_loss": 0.4661687910556793, "eval_runtime": 1489.0076, "eval_samples_per_second": 5.477, "eval_steps_per_second": 0.685, "step": 33000 }, { "epoch": 3.33, "learning_rate": 4.1794865328356706e-05, "loss": 0.2305, "step": 33050 }, { "epoch": 3.33, "learning_rate": 4.178225562392818e-05, "loss": 0.2469, "step": 33100 }, { "epoch": 3.34, "learning_rate": 4.176964591949965e-05, "loss": 0.2559, "step": 33150 }, { "epoch": 3.34, "learning_rate": 4.1757036215071124e-05, "loss": 0.2473, "step": 33200 }, { "epoch": 3.35, "learning_rate": 4.174442651064259e-05, "loss": 0.2336, "step": 33250 }, { "epoch": 3.35, "learning_rate": 4.1731816806214067e-05, "loss": 0.2412, "step": 33300 }, { "epoch": 3.36, "learning_rate": 4.1719207101785535e-05, "loss": 0.2597, "step": 33350 }, { "epoch": 3.36, "learning_rate": 4.1706597397357e-05, "loss": 0.2574, "step": 33400 }, { "epoch": 3.37, "learning_rate": 4.1693987692928484e-05, "loss": 0.2404, "step": 33450 }, { "epoch": 3.37, "learning_rate": 4.168137798849995e-05, "loss": 0.2432, "step": 33500 }, { "epoch": 3.37, "eval_cer": 0.07649487232174547, "eval_loss": 0.4709726870059967, "eval_runtime": 1647.3206, "eval_samples_per_second": 4.95, "eval_steps_per_second": 0.619, "step": 33500 }, { "epoch": 3.38, "learning_rate": 4.166876828407143e-05, "loss": 0.2584, "step": 33550 }, { "epoch": 3.38, "learning_rate": 4.1656158579642895e-05, "loss": 0.278, "step": 33600 }, { "epoch": 3.39, "learning_rate": 4.1643548875214364e-05, "loss": 0.2397, "step": 33650 }, { "epoch": 3.39, "learning_rate": 4.163093917078584e-05, "loss": 0.2406, "step": 33700 }, { "epoch": 3.4, "learning_rate": 4.1618329466357307e-05, "loss": 0.2611, "step": 33750 }, { "epoch": 3.4, "learning_rate": 4.160571976192878e-05, "loss": 0.2394, "step": 33800 }, { "epoch": 3.41, "learning_rate": 4.1593110057500256e-05, "loss": 0.2571, "step": 33850 }, { "epoch": 3.41, "learning_rate": 4.1580500353071724e-05, "loss": 0.2346, "step": 33900 }, { "epoch": 3.42, "learning_rate": 4.15678906486432e-05, "loss": 0.2502, "step": 33950 }, { "epoch": 3.42, "learning_rate": 4.155528094421467e-05, "loss": 0.245, "step": 34000 }, { "epoch": 3.42, "eval_cer": 0.07612438433614158, "eval_loss": 0.46555665135383606, "eval_runtime": 1639.6215, "eval_samples_per_second": 4.974, "eval_steps_per_second": 0.622, "step": 34000 }, { "epoch": 3.43, "learning_rate": 4.154267123978614e-05, "loss": 0.2679, "step": 34050 }, { "epoch": 3.43, "learning_rate": 4.153006153535761e-05, "loss": 0.259, "step": 34100 }, { "epoch": 3.44, "learning_rate": 4.1517451830929085e-05, "loss": 0.2599, "step": 34150 }, { "epoch": 3.44, "learning_rate": 4.150484212650056e-05, "loss": 0.2543, "step": 34200 }, { "epoch": 3.45, "learning_rate": 4.149223242207203e-05, "loss": 0.2519, "step": 34250 }, { "epoch": 3.45, "learning_rate": 4.14796227176435e-05, "loss": 0.2422, "step": 34300 }, { "epoch": 3.46, "learning_rate": 4.146701301321497e-05, "loss": 0.2825, "step": 34350 }, { "epoch": 3.46, "learning_rate": 4.1454403308786446e-05, "loss": 0.2496, "step": 34400 }, { "epoch": 3.47, "learning_rate": 4.1441793604357914e-05, "loss": 0.2539, "step": 34450 }, { "epoch": 3.47, "learning_rate": 4.142918389992939e-05, "loss": 0.2495, "step": 34500 }, { "epoch": 3.47, "eval_cer": 0.077105087827446, "eval_loss": 0.46736568212509155, "eval_runtime": 1511.8312, "eval_samples_per_second": 5.394, "eval_steps_per_second": 0.675, "step": 34500 }, { "epoch": 3.48, "learning_rate": 4.141657419550086e-05, "loss": 0.2626, "step": 34550 }, { "epoch": 3.48, "learning_rate": 4.140396449107233e-05, "loss": 0.2708, "step": 34600 }, { "epoch": 3.49, "learning_rate": 4.139135478664381e-05, "loss": 0.2557, "step": 34650 }, { "epoch": 3.49, "learning_rate": 4.1378745082215275e-05, "loss": 0.2466, "step": 34700 }, { "epoch": 3.5, "learning_rate": 4.136613537778675e-05, "loss": 0.2692, "step": 34750 }, { "epoch": 3.5, "learning_rate": 4.135352567335822e-05, "loss": 0.2596, "step": 34800 }, { "epoch": 3.51, "learning_rate": 4.1340915968929686e-05, "loss": 0.2603, "step": 34850 }, { "epoch": 3.51, "learning_rate": 4.132830626450116e-05, "loss": 0.2624, "step": 34900 }, { "epoch": 3.52, "learning_rate": 4.1315696560072636e-05, "loss": 0.2563, "step": 34950 }, { "epoch": 3.52, "learning_rate": 4.130308685564411e-05, "loss": 0.2589, "step": 35000 }, { "epoch": 3.52, "eval_cer": 0.07430930454112417, "eval_loss": 0.46230170130729675, "eval_runtime": 1642.6651, "eval_samples_per_second": 4.964, "eval_steps_per_second": 0.621, "step": 35000 }, { "epoch": 3.53, "learning_rate": 4.129047715121558e-05, "loss": 0.2578, "step": 35050 }, { "epoch": 3.53, "learning_rate": 4.127786744678705e-05, "loss": 0.2494, "step": 35100 }, { "epoch": 3.54, "learning_rate": 4.126525774235852e-05, "loss": 0.2427, "step": 35150 }, { "epoch": 3.54, "learning_rate": 4.125264803792999e-05, "loss": 0.2669, "step": 35200 }, { "epoch": 3.55, "learning_rate": 4.1240038333501465e-05, "loss": 0.2541, "step": 35250 }, { "epoch": 3.55, "learning_rate": 4.122742862907293e-05, "loss": 0.2609, "step": 35300 }, { "epoch": 3.56, "learning_rate": 4.121481892464441e-05, "loss": 0.2469, "step": 35350 }, { "epoch": 3.56, "learning_rate": 4.120220922021588e-05, "loss": 0.2612, "step": 35400 }, { "epoch": 3.57, "learning_rate": 4.118959951578735e-05, "loss": 0.2665, "step": 35450 }, { "epoch": 3.57, "learning_rate": 4.1176989811358825e-05, "loss": 0.2599, "step": 35500 }, { "epoch": 3.57, "eval_cer": 0.07470469928206278, "eval_loss": 0.46147480607032776, "eval_runtime": 1310.5976, "eval_samples_per_second": 6.222, "eval_steps_per_second": 0.778, "step": 35500 }, { "epoch": 3.58, "learning_rate": 4.1164380106930293e-05, "loss": 0.2611, "step": 35550 }, { "epoch": 3.58, "learning_rate": 4.115177040250177e-05, "loss": 0.2783, "step": 35600 }, { "epoch": 3.59, "learning_rate": 4.1139160698073236e-05, "loss": 0.2695, "step": 35650 }, { "epoch": 3.59, "learning_rate": 4.112655099364471e-05, "loss": 0.2692, "step": 35700 }, { "epoch": 3.6, "learning_rate": 4.1113941289216186e-05, "loss": 0.2551, "step": 35750 }, { "epoch": 3.6, "learning_rate": 4.1101331584787654e-05, "loss": 0.2592, "step": 35800 }, { "epoch": 3.61, "learning_rate": 4.108872188035913e-05, "loss": 0.2572, "step": 35850 }, { "epoch": 3.61, "learning_rate": 4.10761121759306e-05, "loss": 0.2843, "step": 35900 }, { "epoch": 3.62, "learning_rate": 4.106350247150207e-05, "loss": 0.266, "step": 35950 }, { "epoch": 3.62, "learning_rate": 4.105089276707354e-05, "loss": 0.2653, "step": 36000 }, { "epoch": 3.62, "eval_cer": 0.07595003704879856, "eval_loss": 0.45487579703330994, "eval_runtime": 1430.3134, "eval_samples_per_second": 5.702, "eval_steps_per_second": 0.713, "step": 36000 }, { "epoch": 3.63, "learning_rate": 4.103828306264501e-05, "loss": 0.2714, "step": 36050 }, { "epoch": 3.63, "learning_rate": 4.102567335821649e-05, "loss": 0.281, "step": 36100 }, { "epoch": 3.64, "learning_rate": 4.101306365378796e-05, "loss": 0.2784, "step": 36150 }, { "epoch": 3.64, "learning_rate": 4.100045394935943e-05, "loss": 0.2745, "step": 36200 }, { "epoch": 3.65, "learning_rate": 4.09878442449309e-05, "loss": 0.2606, "step": 36250 }, { "epoch": 3.65, "learning_rate": 4.097523454050237e-05, "loss": 0.245, "step": 36300 }, { "epoch": 3.66, "learning_rate": 4.0962624836073844e-05, "loss": 0.2504, "step": 36350 }, { "epoch": 3.66, "learning_rate": 4.095001513164531e-05, "loss": 0.2487, "step": 36400 }, { "epoch": 3.67, "learning_rate": 4.0937405427216794e-05, "loss": 0.2543, "step": 36450 }, { "epoch": 3.67, "learning_rate": 4.092479572278826e-05, "loss": 0.2622, "step": 36500 }, { "epoch": 3.67, "eval_cer": 0.07552039551927471, "eval_loss": 0.45863860845565796, "eval_runtime": 1646.1789, "eval_samples_per_second": 4.954, "eval_steps_per_second": 0.62, "step": 36500 }, { "epoch": 3.68, "learning_rate": 4.091218601835973e-05, "loss": 0.2491, "step": 36550 }, { "epoch": 3.68, "learning_rate": 4.0899576313931205e-05, "loss": 0.2738, "step": 36600 }, { "epoch": 3.69, "learning_rate": 4.088696660950267e-05, "loss": 0.2772, "step": 36650 }, { "epoch": 3.69, "learning_rate": 4.087435690507415e-05, "loss": 0.2571, "step": 36700 }, { "epoch": 3.7, "learning_rate": 4.0861747200645616e-05, "loss": 0.2577, "step": 36750 }, { "epoch": 3.7, "learning_rate": 4.084938969030566e-05, "loss": 0.2597, "step": 36800 }, { "epoch": 3.71, "learning_rate": 4.0836779985877135e-05, "loss": 0.2626, "step": 36850 }, { "epoch": 3.71, "learning_rate": 4.0824170281448604e-05, "loss": 0.2917, "step": 36900 }, { "epoch": 3.72, "learning_rate": 4.081156057702007e-05, "loss": 0.2855, "step": 36950 }, { "epoch": 3.72, "learning_rate": 4.0798950872591546e-05, "loss": 0.2581, "step": 37000 }, { "epoch": 3.72, "eval_cer": 0.07457705216097237, "eval_loss": 0.45668530464172363, "eval_runtime": 1606.7014, "eval_samples_per_second": 5.076, "eval_steps_per_second": 0.635, "step": 37000 }, { "epoch": 3.73, "learning_rate": 4.078634116816302e-05, "loss": 0.26, "step": 37050 }, { "epoch": 3.73, "learning_rate": 4.0773731463734496e-05, "loss": 0.252, "step": 37100 }, { "epoch": 3.74, "learning_rate": 4.0761121759305964e-05, "loss": 0.2662, "step": 37150 }, { "epoch": 3.74, "learning_rate": 4.074851205487743e-05, "loss": 0.2613, "step": 37200 }, { "epoch": 3.75, "learning_rate": 4.073590235044891e-05, "loss": 0.2512, "step": 37250 }, { "epoch": 3.75, "learning_rate": 4.0723292646020375e-05, "loss": 0.2612, "step": 37300 }, { "epoch": 3.76, "learning_rate": 4.071068294159185e-05, "loss": 0.2658, "step": 37350 }, { "epoch": 3.76, "learning_rate": 4.0698325431251895e-05, "loss": 0.2791, "step": 37400 }, { "epoch": 3.77, "learning_rate": 4.068571572682336e-05, "loss": 0.2538, "step": 37450 }, { "epoch": 3.77, "learning_rate": 4.067310602239484e-05, "loss": 0.2705, "step": 37500 }, { "epoch": 3.77, "eval_cer": 0.07581616323887447, "eval_loss": 0.4501054286956787, "eval_runtime": 1633.3689, "eval_samples_per_second": 4.993, "eval_steps_per_second": 0.624, "step": 37500 }, { "epoch": 3.78, "learning_rate": 4.0660496317966306e-05, "loss": 0.2766, "step": 37550 }, { "epoch": 3.78, "learning_rate": 4.064788661353778e-05, "loss": 0.2612, "step": 37600 }, { "epoch": 3.79, "learning_rate": 4.0635276909109256e-05, "loss": 0.2828, "step": 37650 }, { "epoch": 3.79, "learning_rate": 4.0622667204680724e-05, "loss": 0.2586, "step": 37700 }, { "epoch": 3.8, "learning_rate": 4.06100575002522e-05, "loss": 0.2549, "step": 37750 }, { "epoch": 3.8, "learning_rate": 4.059744779582367e-05, "loss": 0.2781, "step": 37800 }, { "epoch": 3.81, "learning_rate": 4.058483809139514e-05, "loss": 0.2639, "step": 37850 }, { "epoch": 3.81, "learning_rate": 4.057222838696661e-05, "loss": 0.2608, "step": 37900 }, { "epoch": 3.82, "learning_rate": 4.055961868253808e-05, "loss": 0.2392, "step": 37950 }, { "epoch": 3.82, "learning_rate": 4.054700897810956e-05, "loss": 0.2575, "step": 38000 }, { "epoch": 3.82, "eval_cer": 0.07289584617587905, "eval_loss": 0.45017656683921814, "eval_runtime": 1306.0998, "eval_samples_per_second": 6.244, "eval_steps_per_second": 0.781, "step": 38000 }, { "epoch": 3.83, "learning_rate": 4.053439927368103e-05, "loss": 0.249, "step": 38050 }, { "epoch": 3.83, "learning_rate": 4.05217895692525e-05, "loss": 0.2633, "step": 38100 }, { "epoch": 3.84, "learning_rate": 4.050917986482397e-05, "loss": 0.2535, "step": 38150 }, { "epoch": 3.84, "learning_rate": 4.049657016039544e-05, "loss": 0.2709, "step": 38200 }, { "epoch": 3.85, "learning_rate": 4.0483960455966914e-05, "loss": 0.2693, "step": 38250 }, { "epoch": 3.85, "learning_rate": 4.047135075153838e-05, "loss": 0.2593, "step": 38300 }, { "epoch": 3.86, "learning_rate": 4.0458741047109857e-05, "loss": 0.2741, "step": 38350 }, { "epoch": 3.86, "learning_rate": 4.044613134268133e-05, "loss": 0.2678, "step": 38400 }, { "epoch": 3.87, "learning_rate": 4.04335216382528e-05, "loss": 0.2625, "step": 38450 }, { "epoch": 3.87, "learning_rate": 4.0420911933824274e-05, "loss": 0.2419, "step": 38500 }, { "epoch": 3.87, "eval_cer": 0.07462063898280812, "eval_loss": 0.44857820868492126, "eval_runtime": 1313.6122, "eval_samples_per_second": 6.208, "eval_steps_per_second": 0.776, "step": 38500 }, { "epoch": 3.88, "learning_rate": 4.040830222939574e-05, "loss": 0.246, "step": 38550 }, { "epoch": 3.88, "learning_rate": 4.039569252496722e-05, "loss": 0.2685, "step": 38600 }, { "epoch": 3.89, "learning_rate": 4.0383082820538685e-05, "loss": 0.2605, "step": 38650 }, { "epoch": 3.89, "learning_rate": 4.037047311611016e-05, "loss": 0.2592, "step": 38700 }, { "epoch": 3.9, "learning_rate": 4.0357863411681635e-05, "loss": 0.2633, "step": 38750 }, { "epoch": 3.9, "learning_rate": 4.03452537072531e-05, "loss": 0.2689, "step": 38800 }, { "epoch": 3.91, "learning_rate": 4.033264400282458e-05, "loss": 0.2479, "step": 38850 }, { "epoch": 3.91, "learning_rate": 4.0320034298396046e-05, "loss": 0.2719, "step": 38900 }, { "epoch": 3.92, "learning_rate": 4.030742459396752e-05, "loss": 0.2657, "step": 38950 }, { "epoch": 3.92, "learning_rate": 4.029481488953899e-05, "loss": 0.2669, "step": 39000 }, { "epoch": 3.92, "eval_cer": 0.07365861555800472, "eval_loss": 0.4459736943244934, "eval_runtime": 1326.0327, "eval_samples_per_second": 6.15, "eval_steps_per_second": 0.769, "step": 39000 }, { "epoch": 3.93, "learning_rate": 4.0282205185110464e-05, "loss": 0.2727, "step": 39050 }, { "epoch": 3.93, "learning_rate": 4.026959548068193e-05, "loss": 0.2476, "step": 39100 }, { "epoch": 3.94, "learning_rate": 4.025698577625341e-05, "loss": 0.2431, "step": 39150 }, { "epoch": 3.94, "learning_rate": 4.024437607182488e-05, "loss": 0.2746, "step": 39200 }, { "epoch": 3.95, "learning_rate": 4.023176636739635e-05, "loss": 0.2638, "step": 39250 }, { "epoch": 3.95, "learning_rate": 4.0219156662967825e-05, "loss": 0.2682, "step": 39300 }, { "epoch": 3.96, "learning_rate": 4.020654695853929e-05, "loss": 0.2664, "step": 39350 }, { "epoch": 3.96, "learning_rate": 4.019393725411076e-05, "loss": 0.2633, "step": 39400 }, { "epoch": 3.97, "learning_rate": 4.0181327549682236e-05, "loss": 0.2707, "step": 39450 }, { "epoch": 3.97, "learning_rate": 4.016871784525371e-05, "loss": 0.2574, "step": 39500 }, { "epoch": 3.97, "eval_cer": 0.07221713709300805, "eval_loss": 0.44468605518341064, "eval_runtime": 1312.9027, "eval_samples_per_second": 6.211, "eval_steps_per_second": 0.777, "step": 39500 }, { "epoch": 3.98, "learning_rate": 4.0156108140825186e-05, "loss": 0.2757, "step": 39550 }, { "epoch": 3.98, "learning_rate": 4.0143498436396654e-05, "loss": 0.2534, "step": 39600 }, { "epoch": 3.99, "learning_rate": 4.013088873196812e-05, "loss": 0.2555, "step": 39650 }, { "epoch": 3.99, "learning_rate": 4.01182790275396e-05, "loss": 0.2537, "step": 39700 }, { "epoch": 4.0, "learning_rate": 4.0105669323111065e-05, "loss": 0.2938, "step": 39750 }, { "epoch": 4.0, "learning_rate": 4.009305961868254e-05, "loss": 0.182, "step": 39800 }, { "epoch": 4.01, "learning_rate": 4.008044991425401e-05, "loss": 0.156, "step": 39850 }, { "epoch": 4.01, "learning_rate": 4.006784020982548e-05, "loss": 0.1644, "step": 39900 }, { "epoch": 4.02, "learning_rate": 4.005523050539696e-05, "loss": 0.1657, "step": 39950 }, { "epoch": 4.02, "learning_rate": 4.0042620800968426e-05, "loss": 0.1693, "step": 40000 }, { "epoch": 4.02, "eval_cer": 0.0748323464031532, "eval_loss": 0.46420931816101074, "eval_runtime": 1316.1933, "eval_samples_per_second": 6.196, "eval_steps_per_second": 0.775, "step": 40000 }, { "epoch": 4.03, "learning_rate": 4.00300110965399e-05, "loss": 0.1751, "step": 40050 }, { "epoch": 4.04, "learning_rate": 4.001740139211137e-05, "loss": 0.1629, "step": 40100 }, { "epoch": 4.04, "learning_rate": 4.0004791687682843e-05, "loss": 0.1705, "step": 40150 }, { "epoch": 4.05, "learning_rate": 3.999218198325431e-05, "loss": 0.1771, "step": 40200 }, { "epoch": 4.05, "learning_rate": 3.9979572278825786e-05, "loss": 0.156, "step": 40250 }, { "epoch": 4.06, "learning_rate": 3.996696257439726e-05, "loss": 0.1758, "step": 40300 }, { "epoch": 4.06, "learning_rate": 3.995435286996873e-05, "loss": 0.1778, "step": 40350 }, { "epoch": 4.07, "learning_rate": 3.9941743165540204e-05, "loss": 0.17, "step": 40400 }, { "epoch": 4.07, "learning_rate": 3.992913346111167e-05, "loss": 0.1849, "step": 40450 }, { "epoch": 4.08, "learning_rate": 3.991652375668315e-05, "loss": 0.1953, "step": 40500 }, { "epoch": 4.08, "eval_cer": 0.07282112590987491, "eval_loss": 0.4596292972564697, "eval_runtime": 1328.5421, "eval_samples_per_second": 6.138, "eval_steps_per_second": 0.768, "step": 40500 }, { "epoch": 4.08, "learning_rate": 3.9903914052254615e-05, "loss": 0.1739, "step": 40550 }, { "epoch": 4.09, "learning_rate": 3.9891304347826083e-05, "loss": 0.189, "step": 40600 }, { "epoch": 4.09, "learning_rate": 3.9878694643397565e-05, "loss": 0.1811, "step": 40650 }, { "epoch": 4.1, "learning_rate": 3.986608493896903e-05, "loss": 0.1771, "step": 40700 }, { "epoch": 4.1, "learning_rate": 3.985347523454051e-05, "loss": 0.1659, "step": 40750 }, { "epoch": 4.11, "learning_rate": 3.9840865530111976e-05, "loss": 0.1954, "step": 40800 }, { "epoch": 4.11, "learning_rate": 3.9828255825683444e-05, "loss": 0.1729, "step": 40850 }, { "epoch": 4.12, "learning_rate": 3.981564612125492e-05, "loss": 0.1664, "step": 40900 }, { "epoch": 4.12, "learning_rate": 3.980303641682639e-05, "loss": 0.1698, "step": 40950 }, { "epoch": 4.13, "learning_rate": 3.979042671239787e-05, "loss": 0.1731, "step": 41000 }, { "epoch": 4.13, "eval_cer": 0.07251913150144147, "eval_loss": 0.46260449290275574, "eval_runtime": 1319.3547, "eval_samples_per_second": 6.181, "eval_steps_per_second": 0.773, "step": 41000 }, { "epoch": 4.13, "learning_rate": 3.977781700796934e-05, "loss": 0.1836, "step": 41050 }, { "epoch": 4.14, "learning_rate": 3.9765207303540805e-05, "loss": 0.1706, "step": 41100 }, { "epoch": 4.14, "learning_rate": 3.975259759911228e-05, "loss": 0.1676, "step": 41150 }, { "epoch": 4.15, "learning_rate": 3.973998789468375e-05, "loss": 0.1662, "step": 41200 }, { "epoch": 4.15, "learning_rate": 3.972737819025522e-05, "loss": 0.1789, "step": 41250 }, { "epoch": 4.16, "learning_rate": 3.971476848582669e-05, "loss": 0.1789, "step": 41300 }, { "epoch": 4.16, "learning_rate": 3.9702158781398166e-05, "loss": 0.1869, "step": 41350 }, { "epoch": 4.17, "learning_rate": 3.968954907696964e-05, "loss": 0.1712, "step": 41400 }, { "epoch": 4.17, "learning_rate": 3.967693937254111e-05, "loss": 0.1755, "step": 41450 }, { "epoch": 4.18, "learning_rate": 3.966458186220115e-05, "loss": 0.1866, "step": 41500 }, { "epoch": 4.18, "eval_cer": 0.0726312119004477, "eval_loss": 0.4616175591945648, "eval_runtime": 1321.8155, "eval_samples_per_second": 6.17, "eval_steps_per_second": 0.772, "step": 41500 }, { "epoch": 4.18, "learning_rate": 3.965197215777262e-05, "loss": 0.1727, "step": 41550 }, { "epoch": 4.19, "learning_rate": 3.9639362453344097e-05, "loss": 0.1671, "step": 41600 }, { "epoch": 4.19, "learning_rate": 3.962675274891557e-05, "loss": 0.1942, "step": 41650 }, { "epoch": 4.2, "learning_rate": 3.961414304448704e-05, "loss": 0.1679, "step": 41700 }, { "epoch": 4.2, "learning_rate": 3.960153334005851e-05, "loss": 0.1989, "step": 41750 }, { "epoch": 4.21, "learning_rate": 3.958892363562998e-05, "loss": 0.1875, "step": 41800 }, { "epoch": 4.21, "learning_rate": 3.957631393120145e-05, "loss": 0.1839, "step": 41850 }, { "epoch": 4.22, "learning_rate": 3.9563704226772925e-05, "loss": 0.1796, "step": 41900 }, { "epoch": 4.22, "learning_rate": 3.95510945223444e-05, "loss": 0.1894, "step": 41950 }, { "epoch": 4.23, "learning_rate": 3.9538484817915875e-05, "loss": 0.1754, "step": 42000 }, { "epoch": 4.23, "eval_cer": 0.07434043798529257, "eval_loss": 0.46099013090133667, "eval_runtime": 1319.4022, "eval_samples_per_second": 6.181, "eval_steps_per_second": 0.773, "step": 42000 }, { "epoch": 4.23, "learning_rate": 3.952587511348734e-05, "loss": 0.1788, "step": 42050 }, { "epoch": 4.24, "learning_rate": 3.951326540905881e-05, "loss": 0.1778, "step": 42100 }, { "epoch": 4.24, "learning_rate": 3.9500655704630286e-05, "loss": 0.1874, "step": 42150 }, { "epoch": 4.25, "learning_rate": 3.9488046000201754e-05, "loss": 0.1902, "step": 42200 }, { "epoch": 4.25, "learning_rate": 3.947543629577323e-05, "loss": 0.1873, "step": 42250 }, { "epoch": 4.26, "learning_rate": 3.94628265913447e-05, "loss": 0.1907, "step": 42300 }, { "epoch": 4.26, "learning_rate": 3.945021688691617e-05, "loss": 0.1826, "step": 42350 }, { "epoch": 4.27, "learning_rate": 3.943760718248765e-05, "loss": 0.18, "step": 42400 }, { "epoch": 4.27, "learning_rate": 3.9424997478059115e-05, "loss": 0.1897, "step": 42450 }, { "epoch": 4.28, "learning_rate": 3.941238777363059e-05, "loss": 0.1898, "step": 42500 }, { "epoch": 4.28, "eval_cer": 0.07309821356297362, "eval_loss": 0.4558318257331848, "eval_runtime": 1318.5584, "eval_samples_per_second": 6.185, "eval_steps_per_second": 0.774, "step": 42500 }, { "epoch": 4.28, "learning_rate": 3.939977806920206e-05, "loss": 0.1962, "step": 42550 }, { "epoch": 4.29, "learning_rate": 3.938716836477353e-05, "loss": 0.1901, "step": 42600 }, { "epoch": 4.29, "learning_rate": 3.9374558660345e-05, "loss": 0.1708, "step": 42650 }, { "epoch": 4.3, "learning_rate": 3.9361948955916476e-05, "loss": 0.196, "step": 42700 }, { "epoch": 4.3, "learning_rate": 3.934933925148795e-05, "loss": 0.1959, "step": 42750 }, { "epoch": 4.31, "learning_rate": 3.933672954705942e-05, "loss": 0.1905, "step": 42800 }, { "epoch": 4.31, "learning_rate": 3.9324119842630894e-05, "loss": 0.1753, "step": 42850 }, { "epoch": 4.32, "learning_rate": 3.931151013820236e-05, "loss": 0.1809, "step": 42900 }, { "epoch": 4.32, "learning_rate": 3.929890043377383e-05, "loss": 0.183, "step": 42950 }, { "epoch": 4.33, "learning_rate": 3.9286290729345305e-05, "loss": 0.1879, "step": 43000 }, { "epoch": 4.33, "eval_cer": 0.07528689468801175, "eval_loss": 0.4561834931373596, "eval_runtime": 1326.0576, "eval_samples_per_second": 6.15, "eval_steps_per_second": 0.769, "step": 43000 }, { "epoch": 4.33, "learning_rate": 3.927368102491677e-05, "loss": 0.1901, "step": 43050 }, { "epoch": 4.34, "learning_rate": 3.9261071320488255e-05, "loss": 0.1844, "step": 43100 }, { "epoch": 4.34, "learning_rate": 3.924846161605972e-05, "loss": 0.1957, "step": 43150 }, { "epoch": 4.35, "learning_rate": 3.923585191163119e-05, "loss": 0.1917, "step": 43200 }, { "epoch": 4.35, "learning_rate": 3.9223242207202666e-05, "loss": 0.1845, "step": 43250 }, { "epoch": 4.36, "learning_rate": 3.9210632502774134e-05, "loss": 0.1786, "step": 43300 }, { "epoch": 4.36, "learning_rate": 3.919802279834561e-05, "loss": 0.1824, "step": 43350 }, { "epoch": 4.37, "learning_rate": 3.918541309391708e-05, "loss": 0.1802, "step": 43400 }, { "epoch": 4.37, "learning_rate": 3.917280338948855e-05, "loss": 0.1948, "step": 43450 }, { "epoch": 4.38, "learning_rate": 3.9160193685060026e-05, "loss": 0.2006, "step": 43500 }, { "epoch": 4.38, "eval_cer": 0.07379248936792882, "eval_loss": 0.45781421661376953, "eval_runtime": 1316.4227, "eval_samples_per_second": 6.195, "eval_steps_per_second": 0.775, "step": 43500 }, { "epoch": 4.38, "learning_rate": 3.9147583980631495e-05, "loss": 0.1907, "step": 43550 }, { "epoch": 4.39, "learning_rate": 3.913497427620297e-05, "loss": 0.1666, "step": 43600 }, { "epoch": 4.39, "learning_rate": 3.912236457177444e-05, "loss": 0.188, "step": 43650 }, { "epoch": 4.4, "learning_rate": 3.910975486734591e-05, "loss": 0.1951, "step": 43700 }, { "epoch": 4.4, "learning_rate": 3.909714516291738e-05, "loss": 0.1942, "step": 43750 }, { "epoch": 4.41, "learning_rate": 3.9084535458488855e-05, "loss": 0.2081, "step": 43800 }, { "epoch": 4.41, "learning_rate": 3.907192575406033e-05, "loss": 0.1903, "step": 43850 }, { "epoch": 4.42, "learning_rate": 3.90593160496318e-05, "loss": 0.1976, "step": 43900 }, { "epoch": 4.42, "learning_rate": 3.904670634520327e-05, "loss": 0.2006, "step": 43950 }, { "epoch": 4.43, "learning_rate": 3.903409664077474e-05, "loss": 0.1792, "step": 44000 }, { "epoch": 4.43, "eval_cer": 0.07377692264584462, "eval_loss": 0.4564515948295593, "eval_runtime": 1339.659, "eval_samples_per_second": 6.087, "eval_steps_per_second": 0.761, "step": 44000 }, { "epoch": 4.43, "learning_rate": 3.9021486936346216e-05, "loss": 0.1969, "step": 44050 }, { "epoch": 4.44, "learning_rate": 3.900912942600626e-05, "loss": 0.1714, "step": 44100 }, { "epoch": 4.44, "learning_rate": 3.899651972157773e-05, "loss": 0.1964, "step": 44150 }, { "epoch": 4.45, "learning_rate": 3.89839100171492e-05, "loss": 0.2029, "step": 44200 }, { "epoch": 4.45, "learning_rate": 3.897130031272067e-05, "loss": 0.1755, "step": 44250 }, { "epoch": 4.46, "learning_rate": 3.895869060829214e-05, "loss": 0.1819, "step": 44300 }, { "epoch": 4.46, "learning_rate": 3.8946333097952185e-05, "loss": 0.2028, "step": 44350 }, { "epoch": 4.47, "learning_rate": 3.893372339352366e-05, "loss": 0.1917, "step": 44400 }, { "epoch": 4.47, "learning_rate": 3.892111368909513e-05, "loss": 0.1935, "step": 44450 }, { "epoch": 4.48, "learning_rate": 3.89085039846666e-05, "loss": 0.1777, "step": 44500 }, { "epoch": 4.48, "eval_cer": 0.07221091040417438, "eval_loss": 0.45735323429107666, "eval_runtime": 1311.9391, "eval_samples_per_second": 6.216, "eval_steps_per_second": 0.777, "step": 44500 }, { "epoch": 4.48, "learning_rate": 3.889589428023807e-05, "loss": 0.21, "step": 44550 }, { "epoch": 4.49, "learning_rate": 3.8883284575809545e-05, "loss": 0.191, "step": 44600 }, { "epoch": 4.49, "learning_rate": 3.887067487138102e-05, "loss": 0.1895, "step": 44650 }, { "epoch": 4.5, "learning_rate": 3.885806516695249e-05, "loss": 0.1843, "step": 44700 }, { "epoch": 4.5, "learning_rate": 3.884545546252396e-05, "loss": 0.1879, "step": 44750 }, { "epoch": 4.51, "learning_rate": 3.883284575809543e-05, "loss": 0.1823, "step": 44800 }, { "epoch": 4.51, "learning_rate": 3.8820236053666906e-05, "loss": 0.184, "step": 44850 }, { "epoch": 4.52, "learning_rate": 3.8807626349238374e-05, "loss": 0.1972, "step": 44900 }, { "epoch": 4.52, "learning_rate": 3.879501664480984e-05, "loss": 0.1989, "step": 44950 }, { "epoch": 4.53, "learning_rate": 3.8782406940381324e-05, "loss": 0.1907, "step": 45000 }, { "epoch": 4.53, "eval_cer": 0.07128313376795621, "eval_loss": 0.45317181944847107, "eval_runtime": 1309.4429, "eval_samples_per_second": 6.228, "eval_steps_per_second": 0.779, "step": 45000 }, { "epoch": 4.53, "learning_rate": 3.876979723595279e-05, "loss": 0.2061, "step": 45050 }, { "epoch": 4.54, "learning_rate": 3.875718753152427e-05, "loss": 0.1835, "step": 45100 }, { "epoch": 4.54, "learning_rate": 3.8744577827095735e-05, "loss": 0.2048, "step": 45150 }, { "epoch": 4.55, "learning_rate": 3.87319681226672e-05, "loss": 0.1868, "step": 45200 }, { "epoch": 4.55, "learning_rate": 3.871935841823868e-05, "loss": 0.2138, "step": 45250 }, { "epoch": 4.56, "learning_rate": 3.8706748713810146e-05, "loss": 0.1951, "step": 45300 }, { "epoch": 4.56, "learning_rate": 3.869413900938162e-05, "loss": 0.2022, "step": 45350 }, { "epoch": 4.57, "learning_rate": 3.8681529304953096e-05, "loss": 0.1872, "step": 45400 }, { "epoch": 4.57, "learning_rate": 3.8668919600524564e-05, "loss": 0.1815, "step": 45450 }, { "epoch": 4.58, "learning_rate": 3.865630989609604e-05, "loss": 0.1831, "step": 45500 }, { "epoch": 4.58, "eval_cer": 0.07090019240468497, "eval_loss": 0.45158177614212036, "eval_runtime": 1320.4616, "eval_samples_per_second": 6.176, "eval_steps_per_second": 0.772, "step": 45500 }, { "epoch": 4.58, "learning_rate": 3.864370019166751e-05, "loss": 0.1903, "step": 45550 }, { "epoch": 4.59, "learning_rate": 3.863109048723898e-05, "loss": 0.1961, "step": 45600 }, { "epoch": 4.59, "learning_rate": 3.861848078281045e-05, "loss": 0.1953, "step": 45650 }, { "epoch": 4.6, "learning_rate": 3.8605871078381925e-05, "loss": 0.1895, "step": 45700 }, { "epoch": 4.6, "learning_rate": 3.85932613739534e-05, "loss": 0.2091, "step": 45750 }, { "epoch": 4.61, "learning_rate": 3.858065166952487e-05, "loss": 0.1787, "step": 45800 }, { "epoch": 4.61, "learning_rate": 3.856804196509634e-05, "loss": 0.2086, "step": 45850 }, { "epoch": 4.62, "learning_rate": 3.855543226066781e-05, "loss": 0.2077, "step": 45900 }, { "epoch": 4.62, "learning_rate": 3.8542822556239286e-05, "loss": 0.1882, "step": 45950 }, { "epoch": 4.63, "learning_rate": 3.8530212851810754e-05, "loss": 0.1864, "step": 46000 }, { "epoch": 4.63, "eval_cer": 0.07166296178681063, "eval_loss": 0.45117127895355225, "eval_runtime": 1317.9389, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 46000 }, { "epoch": 4.63, "learning_rate": 3.851760314738222e-05, "loss": 0.2123, "step": 46050 }, { "epoch": 4.64, "learning_rate": 3.85049934429537e-05, "loss": 0.1898, "step": 46100 }, { "epoch": 4.64, "learning_rate": 3.849238373852517e-05, "loss": 0.1984, "step": 46150 }, { "epoch": 4.65, "learning_rate": 3.8479774034096647e-05, "loss": 0.1895, "step": 46200 }, { "epoch": 4.65, "learning_rate": 3.8467164329668115e-05, "loss": 0.1852, "step": 46250 }, { "epoch": 4.66, "learning_rate": 3.845455462523958e-05, "loss": 0.1835, "step": 46300 }, { "epoch": 4.66, "learning_rate": 3.844194492081106e-05, "loss": 0.2044, "step": 46350 }, { "epoch": 4.67, "learning_rate": 3.8429335216382526e-05, "loss": 0.1875, "step": 46400 }, { "epoch": 4.67, "learning_rate": 3.8416725511954e-05, "loss": 0.1981, "step": 46450 }, { "epoch": 4.68, "learning_rate": 3.8404115807525475e-05, "loss": 0.209, "step": 46500 }, { "epoch": 4.68, "eval_cer": 0.07133917396745933, "eval_loss": 0.44432100653648376, "eval_runtime": 1319.4125, "eval_samples_per_second": 6.181, "eval_steps_per_second": 0.773, "step": 46500 }, { "epoch": 4.68, "learning_rate": 3.839150610309695e-05, "loss": 0.1844, "step": 46550 }, { "epoch": 4.69, "learning_rate": 3.837889639866842e-05, "loss": 0.2, "step": 46600 }, { "epoch": 4.69, "learning_rate": 3.8366286694239886e-05, "loss": 0.2032, "step": 46650 }, { "epoch": 4.7, "learning_rate": 3.835367698981136e-05, "loss": 0.1914, "step": 46700 }, { "epoch": 4.7, "learning_rate": 3.834106728538283e-05, "loss": 0.1992, "step": 46750 }, { "epoch": 4.71, "learning_rate": 3.8328457580954304e-05, "loss": 0.1936, "step": 46800 }, { "epoch": 4.71, "learning_rate": 3.831584787652577e-05, "loss": 0.1897, "step": 46850 }, { "epoch": 4.72, "learning_rate": 3.830323817209725e-05, "loss": 0.2081, "step": 46900 }, { "epoch": 4.72, "learning_rate": 3.829062846766872e-05, "loss": 0.2049, "step": 46950 }, { "epoch": 4.73, "learning_rate": 3.827801876324019e-05, "loss": 0.186, "step": 47000 }, { "epoch": 4.73, "eval_cer": 0.07123020691286994, "eval_loss": 0.4431774914264679, "eval_runtime": 1311.2432, "eval_samples_per_second": 6.219, "eval_steps_per_second": 0.778, "step": 47000 }, { "epoch": 4.73, "learning_rate": 3.8265409058811665e-05, "loss": 0.1919, "step": 47050 }, { "epoch": 4.74, "learning_rate": 3.825279935438313e-05, "loss": 0.1996, "step": 47100 }, { "epoch": 4.74, "learning_rate": 3.824018964995461e-05, "loss": 0.1896, "step": 47150 }, { "epoch": 4.75, "learning_rate": 3.8227579945526076e-05, "loss": 0.2104, "step": 47200 }, { "epoch": 4.75, "learning_rate": 3.821497024109755e-05, "loss": 0.1785, "step": 47250 }, { "epoch": 4.76, "learning_rate": 3.8202360536669026e-05, "loss": 0.1965, "step": 47300 }, { "epoch": 4.76, "learning_rate": 3.8189750832240494e-05, "loss": 0.2034, "step": 47350 }, { "epoch": 4.77, "learning_rate": 3.817714112781197e-05, "loss": 0.2038, "step": 47400 }, { "epoch": 4.77, "learning_rate": 3.816453142338344e-05, "loss": 0.2107, "step": 47450 }, { "epoch": 4.78, "learning_rate": 3.8151921718954905e-05, "loss": 0.1841, "step": 47500 }, { "epoch": 4.78, "eval_cer": 0.07114614661361528, "eval_loss": 0.4510813355445862, "eval_runtime": 1329.955, "eval_samples_per_second": 6.132, "eval_steps_per_second": 0.767, "step": 47500 }, { "epoch": 4.78, "learning_rate": 3.813931201452638e-05, "loss": 0.2018, "step": 47550 }, { "epoch": 4.79, "learning_rate": 3.812670231009785e-05, "loss": 0.1984, "step": 47600 }, { "epoch": 4.79, "learning_rate": 3.811409260566933e-05, "loss": 0.2156, "step": 47650 }, { "epoch": 4.8, "learning_rate": 3.81014829012408e-05, "loss": 0.2163, "step": 47700 }, { "epoch": 4.8, "learning_rate": 3.8088873196812266e-05, "loss": 0.2076, "step": 47750 }, { "epoch": 4.81, "learning_rate": 3.807626349238374e-05, "loss": 0.2151, "step": 47800 }, { "epoch": 4.81, "learning_rate": 3.806365378795521e-05, "loss": 0.1944, "step": 47850 }, { "epoch": 4.82, "learning_rate": 3.8051044083526684e-05, "loss": 0.1868, "step": 47900 }, { "epoch": 4.82, "learning_rate": 3.803843437909815e-05, "loss": 0.1795, "step": 47950 }, { "epoch": 4.83, "learning_rate": 3.802582467466963e-05, "loss": 0.1899, "step": 48000 }, { "epoch": 4.83, "eval_cer": 0.07078188531684507, "eval_loss": 0.44290873408317566, "eval_runtime": 1320.0111, "eval_samples_per_second": 6.178, "eval_steps_per_second": 0.773, "step": 48000 }, { "epoch": 4.83, "learning_rate": 3.80132149702411e-05, "loss": 0.2086, "step": 48050 }, { "epoch": 4.84, "learning_rate": 3.800060526581257e-05, "loss": 0.183, "step": 48100 }, { "epoch": 4.84, "learning_rate": 3.7987995561384045e-05, "loss": 0.2186, "step": 48150 }, { "epoch": 4.85, "learning_rate": 3.797538585695551e-05, "loss": 0.1952, "step": 48200 }, { "epoch": 4.86, "learning_rate": 3.796277615252699e-05, "loss": 0.1961, "step": 48250 }, { "epoch": 4.86, "learning_rate": 3.7950166448098456e-05, "loss": 0.2011, "step": 48300 }, { "epoch": 4.87, "learning_rate": 3.793755674366993e-05, "loss": 0.2014, "step": 48350 }, { "epoch": 4.87, "learning_rate": 3.7924947039241405e-05, "loss": 0.2015, "step": 48400 }, { "epoch": 4.88, "learning_rate": 3.791258952890144e-05, "loss": 0.1985, "step": 48450 }, { "epoch": 4.88, "learning_rate": 3.789997982447291e-05, "loss": 0.2077, "step": 48500 }, { "epoch": 4.88, "eval_cer": 0.07015921643347717, "eval_loss": 0.4386458992958069, "eval_runtime": 1448.8922, "eval_samples_per_second": 5.628, "eval_steps_per_second": 0.704, "step": 48500 }, { "epoch": 4.89, "learning_rate": 3.7887370120044386e-05, "loss": 0.2046, "step": 48550 }, { "epoch": 4.89, "learning_rate": 3.787476041561586e-05, "loss": 0.1944, "step": 48600 }, { "epoch": 4.9, "learning_rate": 3.7862150711187336e-05, "loss": 0.2027, "step": 48650 }, { "epoch": 4.9, "learning_rate": 3.7849541006758804e-05, "loss": 0.2081, "step": 48700 }, { "epoch": 4.91, "learning_rate": 3.783693130233027e-05, "loss": 0.1852, "step": 48750 }, { "epoch": 4.91, "learning_rate": 3.782432159790175e-05, "loss": 0.1873, "step": 48800 }, { "epoch": 4.92, "learning_rate": 3.7811711893473215e-05, "loss": 0.2021, "step": 48850 }, { "epoch": 4.92, "learning_rate": 3.779910218904469e-05, "loss": 0.1997, "step": 48900 }, { "epoch": 4.93, "learning_rate": 3.7786492484616165e-05, "loss": 0.1935, "step": 48950 }, { "epoch": 4.93, "learning_rate": 3.777388278018763e-05, "loss": 0.202, "step": 49000 }, { "epoch": 4.93, "eval_cer": 0.06969532811536809, "eval_loss": 0.4360116124153137, "eval_runtime": 1671.6157, "eval_samples_per_second": 4.879, "eval_steps_per_second": 0.61, "step": 49000 }, { "epoch": 4.94, "learning_rate": 3.776127307575911e-05, "loss": 0.1999, "step": 49050 }, { "epoch": 4.94, "learning_rate": 3.7748663371330576e-05, "loss": 0.2022, "step": 49100 }, { "epoch": 4.95, "learning_rate": 3.773605366690205e-05, "loss": 0.2004, "step": 49150 }, { "epoch": 4.95, "learning_rate": 3.772344396247352e-05, "loss": 0.1937, "step": 49200 }, { "epoch": 4.96, "learning_rate": 3.7710834258044994e-05, "loss": 0.1993, "step": 49250 }, { "epoch": 4.96, "learning_rate": 3.769822455361646e-05, "loss": 0.2098, "step": 49300 }, { "epoch": 4.97, "learning_rate": 3.768561484918794e-05, "loss": 0.1896, "step": 49350 }, { "epoch": 4.97, "learning_rate": 3.767300514475941e-05, "loss": 0.1861, "step": 49400 }, { "epoch": 4.98, "learning_rate": 3.766039544033088e-05, "loss": 0.2015, "step": 49450 }, { "epoch": 4.98, "learning_rate": 3.7647785735902355e-05, "loss": 0.2102, "step": 49500 }, { "epoch": 4.98, "eval_cer": 0.06996307573521629, "eval_loss": 0.44095513224601746, "eval_runtime": 1676.9585, "eval_samples_per_second": 4.863, "eval_steps_per_second": 0.608, "step": 49500 }, { "epoch": 4.99, "learning_rate": 3.763517603147382e-05, "loss": 0.1989, "step": 49550 }, { "epoch": 4.99, "learning_rate": 3.76225663270453e-05, "loss": 0.1952, "step": 49600 }, { "epoch": 5.0, "learning_rate": 3.7609956622616766e-05, "loss": 0.2119, "step": 49650 }, { "epoch": 5.0, "learning_rate": 3.759734691818824e-05, "loss": 0.1822, "step": 49700 }, { "epoch": 5.01, "learning_rate": 3.7584737213759715e-05, "loss": 0.1364, "step": 49750 }, { "epoch": 5.01, "learning_rate": 3.7572127509331183e-05, "loss": 0.1242, "step": 49800 }, { "epoch": 5.02, "learning_rate": 3.755976999899122e-05, "loss": 0.1274, "step": 49850 }, { "epoch": 5.02, "learning_rate": 3.75471602945627e-05, "loss": 0.122, "step": 49900 }, { "epoch": 5.03, "learning_rate": 3.753455059013417e-05, "loss": 0.1227, "step": 49950 }, { "epoch": 5.03, "learning_rate": 3.752194088570564e-05, "loss": 0.121, "step": 50000 }, { "epoch": 5.03, "eval_cer": 0.07083481217193133, "eval_loss": 0.45779597759246826, "eval_runtime": 1672.867, "eval_samples_per_second": 4.875, "eval_steps_per_second": 0.61, "step": 50000 }, { "epoch": 5.04, "learning_rate": 3.7509331181277114e-05, "loss": 0.1427, "step": 50050 }, { "epoch": 5.04, "learning_rate": 3.749672147684858e-05, "loss": 0.1272, "step": 50100 }, { "epoch": 5.05, "learning_rate": 3.748411177242006e-05, "loss": 0.1221, "step": 50150 }, { "epoch": 5.05, "learning_rate": 3.7471502067991525e-05, "loss": 0.1327, "step": 50200 }, { "epoch": 5.06, "learning_rate": 3.7458892363563e-05, "loss": 0.1163, "step": 50250 }, { "epoch": 5.06, "learning_rate": 3.7446282659134475e-05, "loss": 0.1229, "step": 50300 }, { "epoch": 5.07, "learning_rate": 3.743367295470594e-05, "loss": 0.1277, "step": 50350 }, { "epoch": 5.07, "learning_rate": 3.742106325027742e-05, "loss": 0.1352, "step": 50400 }, { "epoch": 5.08, "learning_rate": 3.7408453545848886e-05, "loss": 0.1324, "step": 50450 }, { "epoch": 5.08, "learning_rate": 3.739584384142036e-05, "loss": 0.1384, "step": 50500 }, { "epoch": 5.08, "eval_cer": 0.07126134035703834, "eval_loss": 0.4602069854736328, "eval_runtime": 1705.3714, "eval_samples_per_second": 4.782, "eval_steps_per_second": 0.598, "step": 50500 }, { "epoch": 5.09, "learning_rate": 3.738323413699183e-05, "loss": 0.1385, "step": 50550 }, { "epoch": 5.09, "learning_rate": 3.73706244325633e-05, "loss": 0.1406, "step": 50600 }, { "epoch": 5.1, "learning_rate": 3.735801472813478e-05, "loss": 0.1332, "step": 50650 }, { "epoch": 5.1, "learning_rate": 3.734540502370625e-05, "loss": 0.1295, "step": 50700 }, { "epoch": 5.11, "learning_rate": 3.733279531927772e-05, "loss": 0.1261, "step": 50750 }, { "epoch": 5.11, "learning_rate": 3.732018561484919e-05, "loss": 0.1304, "step": 50800 }, { "epoch": 5.12, "learning_rate": 3.7307575910420665e-05, "loss": 0.1378, "step": 50850 }, { "epoch": 5.12, "learning_rate": 3.729496620599213e-05, "loss": 0.1373, "step": 50900 }, { "epoch": 5.13, "learning_rate": 3.72823565015636e-05, "loss": 0.1206, "step": 50950 }, { "epoch": 5.13, "learning_rate": 3.7269746797135076e-05, "loss": 0.1215, "step": 51000 }, { "epoch": 5.13, "eval_cer": 0.07085971892726604, "eval_loss": 0.45988109707832336, "eval_runtime": 1649.8783, "eval_samples_per_second": 4.943, "eval_steps_per_second": 0.618, "step": 51000 }, { "epoch": 5.14, "learning_rate": 3.725713709270655e-05, "loss": 0.1433, "step": 51050 }, { "epoch": 5.14, "learning_rate": 3.7244527388278025e-05, "loss": 0.1237, "step": 51100 }, { "epoch": 5.15, "learning_rate": 3.7231917683849494e-05, "loss": 0.1359, "step": 51150 }, { "epoch": 5.15, "learning_rate": 3.721930797942096e-05, "loss": 0.1367, "step": 51200 }, { "epoch": 5.16, "learning_rate": 3.7206698274992436e-05, "loss": 0.1426, "step": 51250 }, { "epoch": 5.16, "learning_rate": 3.7194088570563905e-05, "loss": 0.1304, "step": 51300 }, { "epoch": 5.17, "learning_rate": 3.718147886613538e-05, "loss": 0.1261, "step": 51350 }, { "epoch": 5.17, "learning_rate": 3.7168869161706854e-05, "loss": 0.1286, "step": 51400 }, { "epoch": 5.18, "learning_rate": 3.715625945727832e-05, "loss": 0.1301, "step": 51450 }, { "epoch": 5.18, "learning_rate": 3.71436497528498e-05, "loss": 0.1436, "step": 51500 }, { "epoch": 5.18, "eval_cer": 0.07048923094166215, "eval_loss": 0.45715072751045227, "eval_runtime": 1630.7426, "eval_samples_per_second": 5.001, "eval_steps_per_second": 0.625, "step": 51500 }, { "epoch": 5.19, "learning_rate": 3.7131040048421265e-05, "loss": 0.1253, "step": 51550 }, { "epoch": 5.19, "learning_rate": 3.711843034399274e-05, "loss": 0.1306, "step": 51600 }, { "epoch": 5.2, "learning_rate": 3.710582063956421e-05, "loss": 0.1313, "step": 51650 }, { "epoch": 5.2, "learning_rate": 3.709321093513568e-05, "loss": 0.1209, "step": 51700 }, { "epoch": 5.21, "learning_rate": 3.708060123070715e-05, "loss": 0.1325, "step": 51750 }, { "epoch": 5.21, "learning_rate": 3.7067991526278626e-05, "loss": 0.1376, "step": 51800 }, { "epoch": 5.22, "learning_rate": 3.70553818218501e-05, "loss": 0.1517, "step": 51850 }, { "epoch": 5.22, "learning_rate": 3.704277211742157e-05, "loss": 0.141, "step": 51900 }, { "epoch": 5.23, "learning_rate": 3.7030162412993044e-05, "loss": 0.1582, "step": 51950 }, { "epoch": 5.23, "learning_rate": 3.701755270856451e-05, "loss": 0.1267, "step": 52000 }, { "epoch": 5.23, "eval_cer": 0.07566360936244933, "eval_loss": 0.46154525876045227, "eval_runtime": 1362.9575, "eval_samples_per_second": 5.983, "eval_steps_per_second": 0.748, "step": 52000 }, { "epoch": 5.24, "learning_rate": 3.700494300413598e-05, "loss": 0.13, "step": 52050 }, { "epoch": 5.24, "learning_rate": 3.6992333299707455e-05, "loss": 0.1303, "step": 52100 }, { "epoch": 5.25, "learning_rate": 3.697972359527893e-05, "loss": 0.1464, "step": 52150 }, { "epoch": 5.25, "learning_rate": 3.6967113890850405e-05, "loss": 0.1311, "step": 52200 }, { "epoch": 5.26, "learning_rate": 3.695450418642187e-05, "loss": 0.1166, "step": 52250 }, { "epoch": 5.26, "learning_rate": 3.694189448199334e-05, "loss": 0.1519, "step": 52300 }, { "epoch": 5.27, "learning_rate": 3.6929284777564816e-05, "loss": 0.1344, "step": 52350 }, { "epoch": 5.27, "learning_rate": 3.6916675073136284e-05, "loss": 0.1405, "step": 52400 }, { "epoch": 5.28, "learning_rate": 3.690406536870776e-05, "loss": 0.1416, "step": 52450 }, { "epoch": 5.28, "learning_rate": 3.689145566427923e-05, "loss": 0.14, "step": 52500 }, { "epoch": 5.28, "eval_cer": 0.0695303208612756, "eval_loss": 0.4570675790309906, "eval_runtime": 1324.086, "eval_samples_per_second": 6.159, "eval_steps_per_second": 0.77, "step": 52500 }, { "epoch": 5.29, "learning_rate": 3.687909815393927e-05, "loss": 0.1334, "step": 52550 }, { "epoch": 5.29, "learning_rate": 3.6866488449510747e-05, "loss": 0.1507, "step": 52600 }, { "epoch": 5.3, "learning_rate": 3.6853878745082215e-05, "loss": 0.1421, "step": 52650 }, { "epoch": 5.3, "learning_rate": 3.684126904065369e-05, "loss": 0.1269, "step": 52700 }, { "epoch": 5.31, "learning_rate": 3.6828659336225164e-05, "loss": 0.1346, "step": 52750 }, { "epoch": 5.31, "learning_rate": 3.681604963179663e-05, "loss": 0.1337, "step": 52800 }, { "epoch": 5.32, "learning_rate": 3.680343992736811e-05, "loss": 0.1269, "step": 52850 }, { "epoch": 5.32, "learning_rate": 3.6790830222939575e-05, "loss": 0.1438, "step": 52900 }, { "epoch": 5.33, "learning_rate": 3.677822051851105e-05, "loss": 0.1291, "step": 52950 }, { "epoch": 5.33, "learning_rate": 3.676561081408252e-05, "loss": 0.129, "step": 53000 }, { "epoch": 5.33, "eval_cer": 0.07043941743099272, "eval_loss": 0.46095481514930725, "eval_runtime": 1316.8787, "eval_samples_per_second": 6.193, "eval_steps_per_second": 0.775, "step": 53000 }, { "epoch": 5.34, "learning_rate": 3.6753001109653987e-05, "loss": 0.1311, "step": 53050 }, { "epoch": 5.34, "learning_rate": 3.674039140522546e-05, "loss": 0.1603, "step": 53100 }, { "epoch": 5.35, "learning_rate": 3.6727781700796936e-05, "loss": 0.1232, "step": 53150 }, { "epoch": 5.35, "learning_rate": 3.671517199636841e-05, "loss": 0.1366, "step": 53200 }, { "epoch": 5.36, "learning_rate": 3.670256229193988e-05, "loss": 0.1312, "step": 53250 }, { "epoch": 5.36, "learning_rate": 3.668995258751135e-05, "loss": 0.1646, "step": 53300 }, { "epoch": 5.37, "learning_rate": 3.667734288308282e-05, "loss": 0.1448, "step": 53350 }, { "epoch": 5.37, "learning_rate": 3.666473317865429e-05, "loss": 0.1416, "step": 53400 }, { "epoch": 5.38, "learning_rate": 3.6652123474225765e-05, "loss": 0.1374, "step": 53450 }, { "epoch": 5.38, "learning_rate": 3.663951376979724e-05, "loss": 0.1432, "step": 53500 }, { "epoch": 5.38, "eval_cer": 0.07017478315556137, "eval_loss": 0.46022066473960876, "eval_runtime": 1317.1932, "eval_samples_per_second": 6.191, "eval_steps_per_second": 0.774, "step": 53500 }, { "epoch": 5.39, "learning_rate": 3.662690406536871e-05, "loss": 0.1316, "step": 53550 }, { "epoch": 5.39, "learning_rate": 3.661429436094018e-05, "loss": 0.1454, "step": 53600 }, { "epoch": 5.4, "learning_rate": 3.660168465651165e-05, "loss": 0.14, "step": 53650 }, { "epoch": 5.4, "learning_rate": 3.6589074952083126e-05, "loss": 0.1425, "step": 53700 }, { "epoch": 5.41, "learning_rate": 3.6576465247654594e-05, "loss": 0.1476, "step": 53750 }, { "epoch": 5.41, "learning_rate": 3.656385554322607e-05, "loss": 0.1318, "step": 53800 }, { "epoch": 5.42, "learning_rate": 3.655124583879754e-05, "loss": 0.1591, "step": 53850 }, { "epoch": 5.42, "learning_rate": 3.653863613436901e-05, "loss": 0.1431, "step": 53900 }, { "epoch": 5.43, "learning_rate": 3.652602642994049e-05, "loss": 0.1444, "step": 53950 }, { "epoch": 5.43, "learning_rate": 3.6513416725511955e-05, "loss": 0.1429, "step": 54000 }, { "epoch": 5.43, "eval_cer": 0.06953654755010928, "eval_loss": 0.45231354236602783, "eval_runtime": 1316.7646, "eval_samples_per_second": 6.193, "eval_steps_per_second": 0.775, "step": 54000 }, { "epoch": 5.44, "learning_rate": 3.650080702108343e-05, "loss": 0.1552, "step": 54050 }, { "epoch": 5.44, "learning_rate": 3.64881973166549e-05, "loss": 0.1461, "step": 54100 }, { "epoch": 5.45, "learning_rate": 3.647558761222637e-05, "loss": 0.1375, "step": 54150 }, { "epoch": 5.45, "learning_rate": 3.646297790779784e-05, "loss": 0.149, "step": 54200 }, { "epoch": 5.46, "learning_rate": 3.6450368203369316e-05, "loss": 0.141, "step": 54250 }, { "epoch": 5.46, "learning_rate": 3.643775849894079e-05, "loss": 0.1359, "step": 54300 }, { "epoch": 5.47, "learning_rate": 3.642514879451226e-05, "loss": 0.1436, "step": 54350 }, { "epoch": 5.47, "learning_rate": 3.6412539090083734e-05, "loss": 0.1461, "step": 54400 }, { "epoch": 5.48, "learning_rate": 3.63999293856552e-05, "loss": 0.1408, "step": 54450 }, { "epoch": 5.48, "learning_rate": 3.638731968122667e-05, "loss": 0.1379, "step": 54500 }, { "epoch": 5.48, "eval_cer": 0.06990392219129633, "eval_loss": 0.45939967036247253, "eval_runtime": 1315.8943, "eval_samples_per_second": 6.197, "eval_steps_per_second": 0.775, "step": 54500 }, { "epoch": 5.49, "learning_rate": 3.6374709976798145e-05, "loss": 0.1375, "step": 54550 }, { "epoch": 5.49, "learning_rate": 3.636210027236961e-05, "loss": 0.1549, "step": 54600 }, { "epoch": 5.5, "learning_rate": 3.6349490567941094e-05, "loss": 0.131, "step": 54650 }, { "epoch": 5.5, "learning_rate": 3.633688086351256e-05, "loss": 0.1528, "step": 54700 }, { "epoch": 5.51, "learning_rate": 3.632427115908403e-05, "loss": 0.1441, "step": 54750 }, { "epoch": 5.51, "learning_rate": 3.6311661454655505e-05, "loss": 0.1513, "step": 54800 }, { "epoch": 5.52, "learning_rate": 3.6299051750226973e-05, "loss": 0.137, "step": 54850 }, { "epoch": 5.52, "learning_rate": 3.628644204579845e-05, "loss": 0.1465, "step": 54900 }, { "epoch": 5.53, "learning_rate": 3.627408453545849e-05, "loss": 0.1405, "step": 54950 }, { "epoch": 5.53, "learning_rate": 3.626147483102996e-05, "loss": 0.13, "step": 55000 }, { "epoch": 5.53, "eval_cer": 0.07025573011039919, "eval_loss": 0.46031370759010315, "eval_runtime": 1306.6782, "eval_samples_per_second": 6.241, "eval_steps_per_second": 0.781, "step": 55000 }, { "epoch": 5.54, "learning_rate": 3.6248865126601436e-05, "loss": 0.1458, "step": 55050 }, { "epoch": 5.54, "learning_rate": 3.6236255422172904e-05, "loss": 0.1488, "step": 55100 }, { "epoch": 5.55, "learning_rate": 3.622364571774437e-05, "loss": 0.1442, "step": 55150 }, { "epoch": 5.55, "learning_rate": 3.6211036013315854e-05, "loss": 0.1479, "step": 55200 }, { "epoch": 5.56, "learning_rate": 3.619842630888732e-05, "loss": 0.1416, "step": 55250 }, { "epoch": 5.56, "learning_rate": 3.61858166044588e-05, "loss": 0.1399, "step": 55300 }, { "epoch": 5.57, "learning_rate": 3.6173206900030265e-05, "loss": 0.1582, "step": 55350 }, { "epoch": 5.57, "learning_rate": 3.616059719560174e-05, "loss": 0.169, "step": 55400 }, { "epoch": 5.58, "learning_rate": 3.614798749117321e-05, "loss": 0.1434, "step": 55450 }, { "epoch": 5.58, "learning_rate": 3.6135377786744676e-05, "loss": 0.1332, "step": 55500 }, { "epoch": 5.58, "eval_cer": 0.06887963187815616, "eval_loss": 0.4566269814968109, "eval_runtime": 1311.0739, "eval_samples_per_second": 6.22, "eval_steps_per_second": 0.778, "step": 55500 }, { "epoch": 5.59, "learning_rate": 3.612276808231615e-05, "loss": 0.1462, "step": 55550 }, { "epoch": 5.59, "learning_rate": 3.6110158377887626e-05, "loss": 0.1377, "step": 55600 }, { "epoch": 5.6, "learning_rate": 3.60975486734591e-05, "loss": 0.1494, "step": 55650 }, { "epoch": 5.6, "learning_rate": 3.608493896903057e-05, "loss": 0.1422, "step": 55700 }, { "epoch": 5.61, "learning_rate": 3.607232926460204e-05, "loss": 0.1433, "step": 55750 }, { "epoch": 5.61, "learning_rate": 3.605971956017351e-05, "loss": 0.1412, "step": 55800 }, { "epoch": 5.62, "learning_rate": 3.604710985574498e-05, "loss": 0.1378, "step": 55850 }, { "epoch": 5.62, "learning_rate": 3.6034500151316455e-05, "loss": 0.1521, "step": 55900 }, { "epoch": 5.63, "learning_rate": 3.602189044688793e-05, "loss": 0.1633, "step": 55950 }, { "epoch": 5.63, "learning_rate": 3.60092807424594e-05, "loss": 0.1337, "step": 56000 }, { "epoch": 5.63, "eval_cer": 0.06962372119378078, "eval_loss": 0.4581325352191925, "eval_runtime": 1310.9179, "eval_samples_per_second": 6.221, "eval_steps_per_second": 0.778, "step": 56000 }, { "epoch": 5.64, "learning_rate": 3.599667103803087e-05, "loss": 0.1405, "step": 56050 }, { "epoch": 5.64, "learning_rate": 3.598406133360234e-05, "loss": 0.1463, "step": 56100 }, { "epoch": 5.65, "learning_rate": 3.5971451629173815e-05, "loss": 0.1375, "step": 56150 }, { "epoch": 5.66, "learning_rate": 3.5958841924745284e-05, "loss": 0.1555, "step": 56200 }, { "epoch": 5.66, "learning_rate": 3.594623222031676e-05, "loss": 0.1361, "step": 56250 }, { "epoch": 5.67, "learning_rate": 3.5933622515888226e-05, "loss": 0.1449, "step": 56300 }, { "epoch": 5.67, "learning_rate": 3.59210128114597e-05, "loss": 0.1458, "step": 56350 }, { "epoch": 5.68, "learning_rate": 3.5908403107031176e-05, "loss": 0.1436, "step": 56400 }, { "epoch": 5.68, "learning_rate": 3.5895793402602644e-05, "loss": 0.1558, "step": 56450 }, { "epoch": 5.69, "learning_rate": 3.588318369817412e-05, "loss": 0.1402, "step": 56500 }, { "epoch": 5.69, "eval_cer": 0.06995684904638261, "eval_loss": 0.45230379700660706, "eval_runtime": 1319.9539, "eval_samples_per_second": 6.178, "eval_steps_per_second": 0.773, "step": 56500 }, { "epoch": 5.69, "learning_rate": 3.587057399374559e-05, "loss": 0.1399, "step": 56550 }, { "epoch": 5.7, "learning_rate": 3.5857964289317055e-05, "loss": 0.1478, "step": 56600 }, { "epoch": 5.7, "learning_rate": 3.584535458488853e-05, "loss": 0.1526, "step": 56650 }, { "epoch": 5.71, "learning_rate": 3.5832744880460005e-05, "loss": 0.1397, "step": 56700 }, { "epoch": 5.71, "learning_rate": 3.582013517603148e-05, "loss": 0.1448, "step": 56750 }, { "epoch": 5.72, "learning_rate": 3.580752547160295e-05, "loss": 0.1397, "step": 56800 }, { "epoch": 5.72, "learning_rate": 3.5794915767174416e-05, "loss": 0.14, "step": 56850 }, { "epoch": 5.73, "learning_rate": 3.578230606274589e-05, "loss": 0.1598, "step": 56900 }, { "epoch": 5.73, "learning_rate": 3.576969635831736e-05, "loss": 0.1394, "step": 56950 }, { "epoch": 5.74, "learning_rate": 3.5757086653888834e-05, "loss": 0.162, "step": 57000 }, { "epoch": 5.74, "eval_cer": 0.07015921643347717, "eval_loss": 0.45041966438293457, "eval_runtime": 1322.3426, "eval_samples_per_second": 6.167, "eval_steps_per_second": 0.771, "step": 57000 }, { "epoch": 5.74, "learning_rate": 3.57444769494603e-05, "loss": 0.1465, "step": 57050 }, { "epoch": 5.75, "learning_rate": 3.5731867245031784e-05, "loss": 0.1592, "step": 57100 }, { "epoch": 5.75, "learning_rate": 3.571925754060325e-05, "loss": 0.1309, "step": 57150 }, { "epoch": 5.76, "learning_rate": 3.570664783617472e-05, "loss": 0.1414, "step": 57200 }, { "epoch": 5.76, "learning_rate": 3.5694038131746195e-05, "loss": 0.1538, "step": 57250 }, { "epoch": 5.77, "learning_rate": 3.568142842731766e-05, "loss": 0.1499, "step": 57300 }, { "epoch": 5.77, "learning_rate": 3.566881872288914e-05, "loss": 0.1484, "step": 57350 }, { "epoch": 5.78, "learning_rate": 3.5656209018460606e-05, "loss": 0.1451, "step": 57400 }, { "epoch": 5.78, "learning_rate": 3.564359931403208e-05, "loss": 0.1424, "step": 57450 }, { "epoch": 5.79, "learning_rate": 3.5630989609603556e-05, "loss": 0.1465, "step": 57500 }, { "epoch": 5.79, "eval_cer": 0.0694991874171072, "eval_loss": 0.45237475633621216, "eval_runtime": 1334.2508, "eval_samples_per_second": 6.112, "eval_steps_per_second": 0.764, "step": 57500 }, { "epoch": 5.79, "learning_rate": 3.5618379905175024e-05, "loss": 0.1453, "step": 57550 }, { "epoch": 5.8, "learning_rate": 3.56057702007465e-05, "loss": 0.1438, "step": 57600 }, { "epoch": 5.8, "learning_rate": 3.559316049631797e-05, "loss": 0.1555, "step": 57650 }, { "epoch": 5.81, "learning_rate": 3.558055079188944e-05, "loss": 0.1548, "step": 57700 }, { "epoch": 5.81, "learning_rate": 3.556794108746091e-05, "loss": 0.1534, "step": 57750 }, { "epoch": 5.82, "learning_rate": 3.555533138303238e-05, "loss": 0.1461, "step": 57800 }, { "epoch": 5.82, "learning_rate": 3.554272167860386e-05, "loss": 0.1526, "step": 57850 }, { "epoch": 5.83, "learning_rate": 3.553011197417533e-05, "loss": 0.1543, "step": 57900 }, { "epoch": 5.83, "learning_rate": 3.55175022697468e-05, "loss": 0.1317, "step": 57950 }, { "epoch": 5.84, "learning_rate": 3.550489256531827e-05, "loss": 0.1453, "step": 58000 }, { "epoch": 5.84, "eval_cer": 0.06814488259578204, "eval_loss": 0.4510563015937805, "eval_runtime": 1307.847, "eval_samples_per_second": 6.235, "eval_steps_per_second": 0.78, "step": 58000 }, { "epoch": 5.84, "learning_rate": 3.549228286088974e-05, "loss": 0.1614, "step": 58050 }, { "epoch": 5.85, "learning_rate": 3.5479673156461213e-05, "loss": 0.1591, "step": 58100 }, { "epoch": 5.85, "learning_rate": 3.546706345203268e-05, "loss": 0.1445, "step": 58150 }, { "epoch": 5.86, "learning_rate": 3.545445374760416e-05, "loss": 0.1418, "step": 58200 }, { "epoch": 5.86, "learning_rate": 3.544184404317563e-05, "loss": 0.1432, "step": 58250 }, { "epoch": 5.87, "learning_rate": 3.54292343387471e-05, "loss": 0.1538, "step": 58300 }, { "epoch": 5.87, "learning_rate": 3.5416624634318574e-05, "loss": 0.1473, "step": 58350 }, { "epoch": 5.88, "learning_rate": 3.540401492989004e-05, "loss": 0.14, "step": 58400 }, { "epoch": 5.88, "learning_rate": 3.539140522546152e-05, "loss": 0.1692, "step": 58450 }, { "epoch": 5.89, "learning_rate": 3.537904771512156e-05, "loss": 0.1477, "step": 58500 }, { "epoch": 5.89, "eval_cer": 0.07014364971139297, "eval_loss": 0.4529511630535126, "eval_runtime": 1318.8288, "eval_samples_per_second": 6.184, "eval_steps_per_second": 0.773, "step": 58500 }, { "epoch": 5.89, "learning_rate": 3.536643801069303e-05, "loss": 0.152, "step": 58550 }, { "epoch": 5.9, "learning_rate": 3.5353828306264505e-05, "loss": 0.1421, "step": 58600 }, { "epoch": 5.9, "learning_rate": 3.534121860183597e-05, "loss": 0.1541, "step": 58650 }, { "epoch": 5.91, "learning_rate": 3.532860889740745e-05, "loss": 0.1498, "step": 58700 }, { "epoch": 5.91, "learning_rate": 3.5315999192978916e-05, "loss": 0.1535, "step": 58750 }, { "epoch": 5.92, "learning_rate": 3.530338948855039e-05, "loss": 0.1481, "step": 58800 }, { "epoch": 5.92, "learning_rate": 3.5290779784121866e-05, "loss": 0.1638, "step": 58850 }, { "epoch": 5.93, "learning_rate": 3.5278170079693334e-05, "loss": 0.1653, "step": 58900 }, { "epoch": 5.93, "learning_rate": 3.526556037526481e-05, "loss": 0.1547, "step": 58950 }, { "epoch": 5.94, "learning_rate": 3.525295067083628e-05, "loss": 0.1581, "step": 59000 }, { "epoch": 5.94, "eval_cer": 0.07016232977789401, "eval_loss": 0.4454929530620575, "eval_runtime": 1306.5187, "eval_samples_per_second": 6.242, "eval_steps_per_second": 0.781, "step": 59000 }, { "epoch": 5.94, "learning_rate": 3.5240340966407745e-05, "loss": 0.1402, "step": 59050 }, { "epoch": 5.95, "learning_rate": 3.522798345606779e-05, "loss": 0.1755, "step": 59100 }, { "epoch": 5.95, "learning_rate": 3.5215373751639264e-05, "loss": 0.1535, "step": 59150 }, { "epoch": 5.96, "learning_rate": 3.520276404721073e-05, "loss": 0.1555, "step": 59200 }, { "epoch": 5.96, "learning_rate": 3.519015434278221e-05, "loss": 0.1487, "step": 59250 }, { "epoch": 5.97, "learning_rate": 3.5177544638353675e-05, "loss": 0.1496, "step": 59300 }, { "epoch": 5.97, "learning_rate": 3.516493493392515e-05, "loss": 0.1539, "step": 59350 }, { "epoch": 5.98, "learning_rate": 3.5152325229496625e-05, "loss": 0.1732, "step": 59400 }, { "epoch": 5.98, "learning_rate": 3.513971552506809e-05, "loss": 0.1598, "step": 59450 }, { "epoch": 5.99, "learning_rate": 3.512710582063957e-05, "loss": 0.1782, "step": 59500 }, { "epoch": 5.99, "eval_cer": 0.06857452412530589, "eval_loss": 0.44259190559387207, "eval_runtime": 1324.5944, "eval_samples_per_second": 6.157, "eval_steps_per_second": 0.77, "step": 59500 }, { "epoch": 5.99, "learning_rate": 3.5114496116211036e-05, "loss": 0.1502, "step": 59550 }, { "epoch": 6.0, "learning_rate": 3.510188641178251e-05, "loss": 0.1578, "step": 59600 }, { "epoch": 6.0, "learning_rate": 3.508927670735398e-05, "loss": 0.129, "step": 59650 }, { "epoch": 6.01, "learning_rate": 3.507666700292545e-05, "loss": 0.0975, "step": 59700 }, { "epoch": 6.01, "learning_rate": 3.506405729849693e-05, "loss": 0.0814, "step": 59750 }, { "epoch": 6.02, "learning_rate": 3.50514475940684e-05, "loss": 0.086, "step": 59800 }, { "epoch": 6.02, "learning_rate": 3.503883788963987e-05, "loss": 0.0882, "step": 59850 }, { "epoch": 6.03, "learning_rate": 3.502622818521134e-05, "loss": 0.0922, "step": 59900 }, { "epoch": 6.03, "learning_rate": 3.5013618480782815e-05, "loss": 0.0948, "step": 59950 }, { "epoch": 6.04, "learning_rate": 3.500100877635428e-05, "loss": 0.0882, "step": 60000 }, { "epoch": 6.04, "eval_cer": 0.06933729350743155, "eval_loss": 0.46737557649612427, "eval_runtime": 1310.0776, "eval_samples_per_second": 6.225, "eval_steps_per_second": 0.779, "step": 60000 }, { "epoch": 6.04, "learning_rate": 3.498839907192575e-05, "loss": 0.0868, "step": 60050 }, { "epoch": 6.05, "learning_rate": 3.4975789367497226e-05, "loss": 0.0972, "step": 60100 }, { "epoch": 6.05, "learning_rate": 3.49631796630687e-05, "loss": 0.1074, "step": 60150 }, { "epoch": 6.06, "learning_rate": 3.4950569958640176e-05, "loss": 0.0986, "step": 60200 }, { "epoch": 6.06, "learning_rate": 3.4937960254211644e-05, "loss": 0.0943, "step": 60250 }, { "epoch": 6.07, "learning_rate": 3.492535054978311e-05, "loss": 0.0864, "step": 60300 }, { "epoch": 6.07, "learning_rate": 3.491274084535459e-05, "loss": 0.0965, "step": 60350 }, { "epoch": 6.08, "learning_rate": 3.4900131140926055e-05, "loss": 0.0986, "step": 60400 }, { "epoch": 6.08, "learning_rate": 3.488752143649753e-05, "loss": 0.1007, "step": 60450 }, { "epoch": 6.09, "learning_rate": 3.4874911732069005e-05, "loss": 0.0896, "step": 60500 }, { "epoch": 6.09, "eval_cer": 0.0685371639923038, "eval_loss": 0.4645943343639374, "eval_runtime": 1308.155, "eval_samples_per_second": 6.234, "eval_steps_per_second": 0.78, "step": 60500 }, { "epoch": 6.09, "learning_rate": 3.486230202764047e-05, "loss": 0.1023, "step": 60550 }, { "epoch": 6.1, "learning_rate": 3.484969232321195e-05, "loss": 0.0957, "step": 60600 }, { "epoch": 6.1, "learning_rate": 3.4837082618783416e-05, "loss": 0.0898, "step": 60650 }, { "epoch": 6.11, "learning_rate": 3.482447291435489e-05, "loss": 0.1111, "step": 60700 }, { "epoch": 6.11, "learning_rate": 3.481186320992636e-05, "loss": 0.0951, "step": 60750 }, { "epoch": 6.12, "learning_rate": 3.4799253505497834e-05, "loss": 0.1019, "step": 60800 }, { "epoch": 6.12, "learning_rate": 3.47866438010693e-05, "loss": 0.1017, "step": 60850 }, { "epoch": 6.13, "learning_rate": 3.4774034096640776e-05, "loss": 0.0929, "step": 60900 }, { "epoch": 6.13, "learning_rate": 3.476142439221225e-05, "loss": 0.0945, "step": 60950 }, { "epoch": 6.14, "learning_rate": 3.474881468778372e-05, "loss": 0.1105, "step": 61000 }, { "epoch": 6.14, "eval_cer": 0.06832545657195872, "eval_loss": 0.4640914797782898, "eval_runtime": 1308.2307, "eval_samples_per_second": 6.234, "eval_steps_per_second": 0.78, "step": 61000 }, { "epoch": 6.14, "learning_rate": 3.4736204983355194e-05, "loss": 0.1041, "step": 61050 }, { "epoch": 6.15, "learning_rate": 3.472359527892666e-05, "loss": 0.0839, "step": 61100 }, { "epoch": 6.15, "learning_rate": 3.471098557449813e-05, "loss": 0.0988, "step": 61150 }, { "epoch": 6.16, "learning_rate": 3.4698375870069605e-05, "loss": 0.1024, "step": 61200 }, { "epoch": 6.16, "learning_rate": 3.468576616564108e-05, "loss": 0.1065, "step": 61250 }, { "epoch": 6.17, "learning_rate": 3.4673156461212555e-05, "loss": 0.104, "step": 61300 }, { "epoch": 6.17, "learning_rate": 3.466054675678402e-05, "loss": 0.1027, "step": 61350 }, { "epoch": 6.18, "learning_rate": 3.464793705235549e-05, "loss": 0.1039, "step": 61400 }, { "epoch": 6.18, "learning_rate": 3.4635327347926966e-05, "loss": 0.1141, "step": 61450 }, { "epoch": 6.19, "learning_rate": 3.4622717643498434e-05, "loss": 0.0991, "step": 61500 }, { "epoch": 6.19, "eval_cer": 0.06922209976400849, "eval_loss": 0.4651162922382355, "eval_runtime": 1325.5484, "eval_samples_per_second": 6.152, "eval_steps_per_second": 0.769, "step": 61500 }, { "epoch": 6.19, "learning_rate": 3.461010793906991e-05, "loss": 0.1054, "step": 61550 }, { "epoch": 6.2, "learning_rate": 3.459749823464138e-05, "loss": 0.0926, "step": 61600 }, { "epoch": 6.2, "learning_rate": 3.458488853021286e-05, "loss": 0.1051, "step": 61650 }, { "epoch": 6.21, "learning_rate": 3.457227882578433e-05, "loss": 0.0965, "step": 61700 }, { "epoch": 6.21, "learning_rate": 3.4559669121355795e-05, "loss": 0.0938, "step": 61750 }, { "epoch": 6.22, "learning_rate": 3.454705941692727e-05, "loss": 0.1084, "step": 61800 }, { "epoch": 6.22, "learning_rate": 3.453444971249874e-05, "loss": 0.1068, "step": 61850 }, { "epoch": 6.23, "learning_rate": 3.452184000807021e-05, "loss": 0.1069, "step": 61900 }, { "epoch": 6.23, "learning_rate": 3.450923030364168e-05, "loss": 0.0932, "step": 61950 }, { "epoch": 6.24, "learning_rate": 3.4496620599213156e-05, "loss": 0.0984, "step": 62000 }, { "epoch": 6.24, "eval_cer": 0.06951786748360825, "eval_loss": 0.4679470956325531, "eval_runtime": 1321.4584, "eval_samples_per_second": 6.171, "eval_steps_per_second": 0.772, "step": 62000 }, { "epoch": 6.24, "learning_rate": 3.448401089478463e-05, "loss": 0.1066, "step": 62050 }, { "epoch": 6.25, "learning_rate": 3.44714011903561e-05, "loss": 0.092, "step": 62100 }, { "epoch": 6.25, "learning_rate": 3.4458791485927574e-05, "loss": 0.1087, "step": 62150 }, { "epoch": 6.26, "learning_rate": 3.444618178149904e-05, "loss": 0.0997, "step": 62200 }, { "epoch": 6.26, "learning_rate": 3.443357207707052e-05, "loss": 0.1018, "step": 62250 }, { "epoch": 6.27, "learning_rate": 3.4420962372641985e-05, "loss": 0.1114, "step": 62300 }, { "epoch": 6.27, "learning_rate": 3.440835266821345e-05, "loss": 0.1029, "step": 62350 }, { "epoch": 6.28, "learning_rate": 3.4395742963784935e-05, "loss": 0.0982, "step": 62400 }, { "epoch": 6.28, "learning_rate": 3.43831332593564e-05, "loss": 0.106, "step": 62450 }, { "epoch": 6.29, "learning_rate": 3.437052355492788e-05, "loss": 0.1055, "step": 62500 }, { "epoch": 6.29, "eval_cer": 0.06921587307517481, "eval_loss": 0.46069595217704773, "eval_runtime": 1313.3714, "eval_samples_per_second": 6.209, "eval_steps_per_second": 0.777, "step": 62500 }, { "epoch": 6.29, "learning_rate": 3.4357913850499346e-05, "loss": 0.1087, "step": 62550 }, { "epoch": 6.3, "learning_rate": 3.4345304146070814e-05, "loss": 0.1005, "step": 62600 }, { "epoch": 6.3, "learning_rate": 3.433269444164229e-05, "loss": 0.1137, "step": 62650 }, { "epoch": 6.31, "learning_rate": 3.432008473721376e-05, "loss": 0.1117, "step": 62700 }, { "epoch": 6.31, "learning_rate": 3.430747503278524e-05, "loss": 0.107, "step": 62750 }, { "epoch": 6.32, "learning_rate": 3.4294865328356706e-05, "loss": 0.1053, "step": 62800 }, { "epoch": 6.32, "learning_rate": 3.4282255623928175e-05, "loss": 0.0974, "step": 62850 }, { "epoch": 6.33, "learning_rate": 3.426964591949965e-05, "loss": 0.0974, "step": 62900 }, { "epoch": 6.33, "learning_rate": 3.425703621507112e-05, "loss": 0.1036, "step": 62950 }, { "epoch": 6.34, "learning_rate": 3.424442651064259e-05, "loss": 0.1021, "step": 63000 }, { "epoch": 6.34, "eval_cer": 0.0683628167049608, "eval_loss": 0.46667516231536865, "eval_runtime": 1319.0468, "eval_samples_per_second": 6.182, "eval_steps_per_second": 0.773, "step": 63000 }, { "epoch": 6.34, "learning_rate": 3.423181680621406e-05, "loss": 0.1098, "step": 63050 }, { "epoch": 6.35, "learning_rate": 3.4219207101785535e-05, "loss": 0.1036, "step": 63100 }, { "epoch": 6.35, "learning_rate": 3.420659739735701e-05, "loss": 0.0974, "step": 63150 }, { "epoch": 6.36, "learning_rate": 3.419398769292848e-05, "loss": 0.1115, "step": 63200 }, { "epoch": 6.36, "learning_rate": 3.418137798849995e-05, "loss": 0.1118, "step": 63250 }, { "epoch": 6.37, "learning_rate": 3.416876828407142e-05, "loss": 0.1062, "step": 63300 }, { "epoch": 6.37, "learning_rate": 3.4156158579642896e-05, "loss": 0.1017, "step": 63350 }, { "epoch": 6.38, "learning_rate": 3.4143548875214364e-05, "loss": 0.1041, "step": 63400 }, { "epoch": 6.38, "learning_rate": 3.413093917078584e-05, "loss": 0.102, "step": 63450 }, { "epoch": 6.39, "learning_rate": 3.4118581660445884e-05, "loss": 0.109, "step": 63500 }, { "epoch": 6.39, "eval_cer": 0.06823516958387038, "eval_loss": 0.4620670676231384, "eval_runtime": 1316.2414, "eval_samples_per_second": 6.196, "eval_steps_per_second": 0.775, "step": 63500 }, { "epoch": 6.39, "learning_rate": 3.410597195601735e-05, "loss": 0.1135, "step": 63550 }, { "epoch": 6.4, "learning_rate": 3.409336225158882e-05, "loss": 0.1051, "step": 63600 }, { "epoch": 6.4, "learning_rate": 3.4080752547160295e-05, "loss": 0.1067, "step": 63650 }, { "epoch": 6.41, "learning_rate": 3.406814284273177e-05, "loss": 0.0946, "step": 63700 }, { "epoch": 6.41, "learning_rate": 3.4055533138303245e-05, "loss": 0.1045, "step": 63750 }, { "epoch": 6.42, "learning_rate": 3.404292343387471e-05, "loss": 0.1011, "step": 63800 }, { "epoch": 6.42, "learning_rate": 3.403031372944618e-05, "loss": 0.0985, "step": 63850 }, { "epoch": 6.43, "learning_rate": 3.4017704025017656e-05, "loss": 0.096, "step": 63900 }, { "epoch": 6.43, "learning_rate": 3.4005094320589124e-05, "loss": 0.1058, "step": 63950 }, { "epoch": 6.44, "learning_rate": 3.39924846161606e-05, "loss": 0.0999, "step": 64000 }, { "epoch": 6.44, "eval_cer": 0.06716417910447761, "eval_loss": 0.46670693159103394, "eval_runtime": 1372.007, "eval_samples_per_second": 5.944, "eval_steps_per_second": 0.743, "step": 64000 }, { "epoch": 6.44, "learning_rate": 3.397987491173207e-05, "loss": 0.1036, "step": 64050 }, { "epoch": 6.45, "learning_rate": 3.396726520730354e-05, "loss": 0.1112, "step": 64100 }, { "epoch": 6.45, "learning_rate": 3.3954655502875016e-05, "loss": 0.1031, "step": 64150 }, { "epoch": 6.46, "learning_rate": 3.3942045798446485e-05, "loss": 0.0939, "step": 64200 }, { "epoch": 6.47, "learning_rate": 3.392943609401796e-05, "loss": 0.1052, "step": 64250 }, { "epoch": 6.47, "learning_rate": 3.391682638958943e-05, "loss": 0.1114, "step": 64300 }, { "epoch": 6.48, "learning_rate": 3.39042166851609e-05, "loss": 0.1191, "step": 64350 }, { "epoch": 6.48, "learning_rate": 3.389160698073237e-05, "loss": 0.1061, "step": 64400 }, { "epoch": 6.49, "learning_rate": 3.3878997276303845e-05, "loss": 0.1223, "step": 64450 }, { "epoch": 6.49, "learning_rate": 3.386638757187532e-05, "loss": 0.1077, "step": 64500 }, { "epoch": 6.49, "eval_cer": 0.06788647500918436, "eval_loss": 0.45837146043777466, "eval_runtime": 1682.0297, "eval_samples_per_second": 4.848, "eval_steps_per_second": 0.606, "step": 64500 }, { "epoch": 6.5, "learning_rate": 3.385377786744679e-05, "loss": 0.1076, "step": 64550 }, { "epoch": 6.5, "learning_rate": 3.384116816301826e-05, "loss": 0.1055, "step": 64600 }, { "epoch": 6.51, "learning_rate": 3.382855845858973e-05, "loss": 0.0976, "step": 64650 }, { "epoch": 6.51, "learning_rate": 3.3815948754161206e-05, "loss": 0.1102, "step": 64700 }, { "epoch": 6.52, "learning_rate": 3.3803339049732674e-05, "loss": 0.099, "step": 64750 }, { "epoch": 6.52, "learning_rate": 3.379072934530414e-05, "loss": 0.1042, "step": 64800 }, { "epoch": 6.53, "learning_rate": 3.3778119640875624e-05, "loss": 0.0986, "step": 64850 }, { "epoch": 6.53, "learning_rate": 3.376550993644709e-05, "loss": 0.0996, "step": 64900 }, { "epoch": 6.54, "learning_rate": 3.375290023201857e-05, "loss": 0.1087, "step": 64950 }, { "epoch": 6.54, "learning_rate": 3.3740290527590035e-05, "loss": 0.1117, "step": 65000 }, { "epoch": 6.54, "eval_cer": 0.06811063580719681, "eval_loss": 0.4665793180465698, "eval_runtime": 1646.839, "eval_samples_per_second": 4.952, "eval_steps_per_second": 0.619, "step": 65000 }, { "epoch": 6.55, "learning_rate": 3.37276808231615e-05, "loss": 0.1117, "step": 65050 }, { "epoch": 6.55, "learning_rate": 3.371507111873298e-05, "loss": 0.0999, "step": 65100 }, { "epoch": 6.56, "learning_rate": 3.3702461414304446e-05, "loss": 0.1099, "step": 65150 }, { "epoch": 6.56, "learning_rate": 3.368985170987593e-05, "loss": 0.1258, "step": 65200 }, { "epoch": 6.57, "learning_rate": 3.3677242005447396e-05, "loss": 0.1246, "step": 65250 }, { "epoch": 6.57, "learning_rate": 3.3664632301018864e-05, "loss": 0.116, "step": 65300 }, { "epoch": 6.58, "learning_rate": 3.365202259659034e-05, "loss": 0.1156, "step": 65350 }, { "epoch": 6.58, "learning_rate": 3.363941289216181e-05, "loss": 0.1107, "step": 65400 }, { "epoch": 6.59, "learning_rate": 3.362680318773328e-05, "loss": 0.1245, "step": 65450 }, { "epoch": 6.59, "learning_rate": 3.361419348330475e-05, "loss": 0.1107, "step": 65500 }, { "epoch": 6.59, "eval_cer": 0.06765608752233825, "eval_loss": 0.46230730414390564, "eval_runtime": 1681.8092, "eval_samples_per_second": 4.849, "eval_steps_per_second": 0.606, "step": 65500 }, { "epoch": 6.6, "learning_rate": 3.3601583778876225e-05, "loss": 0.1148, "step": 65550 }, { "epoch": 6.6, "learning_rate": 3.35889740744477e-05, "loss": 0.1049, "step": 65600 }, { "epoch": 6.61, "learning_rate": 3.357636437001917e-05, "loss": 0.1072, "step": 65650 }, { "epoch": 6.61, "learning_rate": 3.356375466559064e-05, "loss": 0.1099, "step": 65700 }, { "epoch": 6.62, "learning_rate": 3.355114496116211e-05, "loss": 0.097, "step": 65750 }, { "epoch": 6.62, "learning_rate": 3.3538535256733586e-05, "loss": 0.1105, "step": 65800 }, { "epoch": 6.63, "learning_rate": 3.3525925552305054e-05, "loss": 0.1061, "step": 65850 }, { "epoch": 6.63, "learning_rate": 3.351331584787653e-05, "loss": 0.1023, "step": 65900 }, { "epoch": 6.64, "learning_rate": 3.3500706143448003e-05, "loss": 0.1026, "step": 65950 }, { "epoch": 6.64, "learning_rate": 3.348809643901947e-05, "loss": 0.1187, "step": 66000 }, { "epoch": 6.64, "eval_cer": 0.06827875640570613, "eval_loss": 0.45917195081710815, "eval_runtime": 1615.7583, "eval_samples_per_second": 5.047, "eval_steps_per_second": 0.631, "step": 66000 }, { "epoch": 6.65, "learning_rate": 3.3475486734590946e-05, "loss": 0.1096, "step": 66050 }, { "epoch": 6.65, "learning_rate": 3.3462877030162414e-05, "loss": 0.1151, "step": 66100 }, { "epoch": 6.66, "learning_rate": 3.345026732573389e-05, "loss": 0.1112, "step": 66150 }, { "epoch": 6.66, "learning_rate": 3.343765762130536e-05, "loss": 0.1089, "step": 66200 }, { "epoch": 6.67, "learning_rate": 3.3425047916876826e-05, "loss": 0.1062, "step": 66250 }, { "epoch": 6.67, "learning_rate": 3.34124382124483e-05, "loss": 0.1004, "step": 66300 }, { "epoch": 6.68, "learning_rate": 3.3399828508019775e-05, "loss": 0.1068, "step": 66350 }, { "epoch": 6.68, "learning_rate": 3.338721880359125e-05, "loss": 0.1084, "step": 66400 }, { "epoch": 6.69, "learning_rate": 3.337460909916272e-05, "loss": 0.1101, "step": 66450 }, { "epoch": 6.69, "learning_rate": 3.3361999394734186e-05, "loss": 0.1079, "step": 66500 }, { "epoch": 6.69, "eval_cer": 0.06767476758883928, "eval_loss": 0.4588070511817932, "eval_runtime": 1553.9147, "eval_samples_per_second": 5.248, "eval_steps_per_second": 0.656, "step": 66500 }, { "epoch": 6.7, "learning_rate": 3.334938969030566e-05, "loss": 0.1081, "step": 66550 }, { "epoch": 6.7, "learning_rate": 3.333677998587713e-05, "loss": 0.1246, "step": 66600 }, { "epoch": 6.71, "learning_rate": 3.3324170281448604e-05, "loss": 0.1059, "step": 66650 }, { "epoch": 6.71, "learning_rate": 3.331156057702008e-05, "loss": 0.1157, "step": 66700 }, { "epoch": 6.72, "learning_rate": 3.329895087259155e-05, "loss": 0.1143, "step": 66750 }, { "epoch": 6.72, "learning_rate": 3.328659336225159e-05, "loss": 0.1027, "step": 66800 }, { "epoch": 6.73, "learning_rate": 3.327398365782306e-05, "loss": 0.1214, "step": 66850 }, { "epoch": 6.73, "learning_rate": 3.3261373953394535e-05, "loss": 0.1089, "step": 66900 }, { "epoch": 6.74, "learning_rate": 3.324876424896601e-05, "loss": 0.1145, "step": 66950 }, { "epoch": 6.74, "learning_rate": 3.323615454453748e-05, "loss": 0.1125, "step": 67000 }, { "epoch": 6.74, "eval_cer": 0.06958324771636187, "eval_loss": 0.4630543887615204, "eval_runtime": 1528.4328, "eval_samples_per_second": 5.336, "eval_steps_per_second": 0.667, "step": 67000 }, { "epoch": 6.75, "learning_rate": 3.322354484010895e-05, "loss": 0.1238, "step": 67050 }, { "epoch": 6.75, "learning_rate": 3.321093513568042e-05, "loss": 0.1118, "step": 67100 }, { "epoch": 6.76, "learning_rate": 3.319832543125189e-05, "loss": 0.1222, "step": 67150 }, { "epoch": 6.76, "learning_rate": 3.3185715726823364e-05, "loss": 0.1136, "step": 67200 }, { "epoch": 6.77, "learning_rate": 3.317310602239483e-05, "loss": 0.1124, "step": 67250 }, { "epoch": 6.77, "learning_rate": 3.3160496317966313e-05, "loss": 0.1126, "step": 67300 }, { "epoch": 6.78, "learning_rate": 3.314788661353778e-05, "loss": 0.1028, "step": 67350 }, { "epoch": 6.78, "learning_rate": 3.313527690910925e-05, "loss": 0.1035, "step": 67400 }, { "epoch": 6.79, "learning_rate": 3.3122667204680725e-05, "loss": 0.1062, "step": 67450 }, { "epoch": 6.79, "learning_rate": 3.311005750025219e-05, "loss": 0.1077, "step": 67500 }, { "epoch": 6.79, "eval_cer": 0.06730116625881855, "eval_loss": 0.45617401599884033, "eval_runtime": 1312.6867, "eval_samples_per_second": 6.212, "eval_steps_per_second": 0.777, "step": 67500 }, { "epoch": 6.8, "learning_rate": 3.309744779582367e-05, "loss": 0.1165, "step": 67550 }, { "epoch": 6.8, "learning_rate": 3.3084838091395136e-05, "loss": 0.1151, "step": 67600 }, { "epoch": 6.81, "learning_rate": 3.307222838696662e-05, "loss": 0.108, "step": 67650 }, { "epoch": 6.81, "learning_rate": 3.3059618682538085e-05, "loss": 0.1111, "step": 67700 }, { "epoch": 6.82, "learning_rate": 3.3047008978109553e-05, "loss": 0.126, "step": 67750 }, { "epoch": 6.82, "learning_rate": 3.303439927368103e-05, "loss": 0.1104, "step": 67800 }, { "epoch": 6.83, "learning_rate": 3.3021789569252496e-05, "loss": 0.1074, "step": 67850 }, { "epoch": 6.83, "learning_rate": 3.300917986482397e-05, "loss": 0.1075, "step": 67900 }, { "epoch": 6.84, "learning_rate": 3.299657016039544e-05, "loss": 0.1045, "step": 67950 }, { "epoch": 6.84, "learning_rate": 3.2983960455966914e-05, "loss": 0.1173, "step": 68000 }, { "epoch": 6.84, "eval_cer": 0.0686461310468932, "eval_loss": 0.46393048763275146, "eval_runtime": 1297.1193, "eval_samples_per_second": 6.287, "eval_steps_per_second": 0.786, "step": 68000 }, { "epoch": 6.85, "learning_rate": 3.297135075153839e-05, "loss": 0.1082, "step": 68050 }, { "epoch": 6.85, "learning_rate": 3.295874104710986e-05, "loss": 0.1099, "step": 68100 }, { "epoch": 6.86, "learning_rate": 3.294613134268133e-05, "loss": 0.1145, "step": 68150 }, { "epoch": 6.86, "learning_rate": 3.29335216382528e-05, "loss": 0.103, "step": 68200 }, { "epoch": 6.87, "learning_rate": 3.2920911933824275e-05, "loss": 0.1073, "step": 68250 }, { "epoch": 6.87, "learning_rate": 3.290830222939574e-05, "loss": 0.112, "step": 68300 }, { "epoch": 6.88, "learning_rate": 3.289569252496721e-05, "loss": 0.117, "step": 68350 }, { "epoch": 6.88, "learning_rate": 3.2883082820538686e-05, "loss": 0.1091, "step": 68400 }, { "epoch": 6.89, "learning_rate": 3.287047311611016e-05, "loss": 0.1172, "step": 68450 }, { "epoch": 6.89, "learning_rate": 3.2857863411681636e-05, "loss": 0.1129, "step": 68500 }, { "epoch": 6.89, "eval_cer": 0.06727314615906699, "eval_loss": 0.45872893929481506, "eval_runtime": 1321.4497, "eval_samples_per_second": 6.171, "eval_steps_per_second": 0.772, "step": 68500 }, { "epoch": 6.9, "learning_rate": 3.2845253707253104e-05, "loss": 0.1234, "step": 68550 }, { "epoch": 6.9, "learning_rate": 3.283264400282457e-05, "loss": 0.1164, "step": 68600 }, { "epoch": 6.91, "learning_rate": 3.282003429839605e-05, "loss": 0.1301, "step": 68650 }, { "epoch": 6.91, "learning_rate": 3.2807424593967515e-05, "loss": 0.1132, "step": 68700 }, { "epoch": 6.92, "learning_rate": 3.279481488953899e-05, "loss": 0.1087, "step": 68750 }, { "epoch": 6.92, "learning_rate": 3.2782205185110465e-05, "loss": 0.1097, "step": 68800 }, { "epoch": 6.93, "learning_rate": 3.276959548068193e-05, "loss": 0.127, "step": 68850 }, { "epoch": 6.93, "learning_rate": 3.275698577625341e-05, "loss": 0.1069, "step": 68900 }, { "epoch": 6.94, "learning_rate": 3.2744376071824876e-05, "loss": 0.1218, "step": 68950 }, { "epoch": 6.94, "learning_rate": 3.273176636739635e-05, "loss": 0.1036, "step": 69000 }, { "epoch": 6.94, "eval_cer": 0.06748485357941207, "eval_loss": 0.45707839727401733, "eval_runtime": 1315.0237, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 69000 }, { "epoch": 6.95, "learning_rate": 3.271915666296782e-05, "loss": 0.1057, "step": 69050 }, { "epoch": 6.95, "learning_rate": 3.2706546958539294e-05, "loss": 0.1193, "step": 69100 }, { "epoch": 6.96, "learning_rate": 3.269393725411076e-05, "loss": 0.1066, "step": 69150 }, { "epoch": 6.96, "learning_rate": 3.2681327549682237e-05, "loss": 0.1143, "step": 69200 }, { "epoch": 6.97, "learning_rate": 3.266871784525371e-05, "loss": 0.1237, "step": 69250 }, { "epoch": 6.97, "learning_rate": 3.265610814082518e-05, "loss": 0.1104, "step": 69300 }, { "epoch": 6.98, "learning_rate": 3.2643498436396654e-05, "loss": 0.1098, "step": 69350 }, { "epoch": 6.98, "learning_rate": 3.263088873196812e-05, "loss": 0.1136, "step": 69400 }, { "epoch": 6.99, "learning_rate": 3.26182790275396e-05, "loss": 0.1103, "step": 69450 }, { "epoch": 6.99, "learning_rate": 3.2605669323111066e-05, "loss": 0.1147, "step": 69500 }, { "epoch": 6.99, "eval_cer": 0.06715172572681026, "eval_loss": 0.4610083997249603, "eval_runtime": 1314.2566, "eval_samples_per_second": 6.205, "eval_steps_per_second": 0.776, "step": 69500 }, { "epoch": 7.0, "learning_rate": 3.259305961868254e-05, "loss": 0.1113, "step": 69550 }, { "epoch": 7.0, "learning_rate": 3.2580449914254015e-05, "loss": 0.0821, "step": 69600 }, { "epoch": 7.01, "learning_rate": 3.256784020982548e-05, "loss": 0.0645, "step": 69650 }, { "epoch": 7.01, "learning_rate": 3.255523050539696e-05, "loss": 0.0616, "step": 69700 }, { "epoch": 7.02, "learning_rate": 3.2542620800968426e-05, "loss": 0.0687, "step": 69750 }, { "epoch": 7.02, "learning_rate": 3.2530011096539894e-05, "loss": 0.0644, "step": 69800 }, { "epoch": 7.03, "learning_rate": 3.251740139211137e-05, "loss": 0.0646, "step": 69850 }, { "epoch": 7.03, "learning_rate": 3.250479168768284e-05, "loss": 0.0749, "step": 69900 }, { "epoch": 7.04, "learning_rate": 3.249218198325432e-05, "loss": 0.0798, "step": 69950 }, { "epoch": 7.04, "learning_rate": 3.247957227882579e-05, "loss": 0.0676, "step": 70000 }, { "epoch": 7.04, "eval_cer": 0.06708323214963978, "eval_loss": 0.4726824164390564, "eval_runtime": 1309.2974, "eval_samples_per_second": 6.229, "eval_steps_per_second": 0.779, "step": 70000 }, { "epoch": 7.05, "learning_rate": 3.2466962574397255e-05, "loss": 0.0793, "step": 70050 }, { "epoch": 7.05, "learning_rate": 3.245435286996873e-05, "loss": 0.0641, "step": 70100 }, { "epoch": 7.06, "learning_rate": 3.24417431655402e-05, "loss": 0.0771, "step": 70150 }, { "epoch": 7.06, "learning_rate": 3.242913346111167e-05, "loss": 0.0718, "step": 70200 }, { "epoch": 7.07, "learning_rate": 3.241652375668314e-05, "loss": 0.0746, "step": 70250 }, { "epoch": 7.07, "learning_rate": 3.2403914052254616e-05, "loss": 0.0693, "step": 70300 }, { "epoch": 7.08, "learning_rate": 3.239130434782609e-05, "loss": 0.0745, "step": 70350 }, { "epoch": 7.08, "learning_rate": 3.237869464339756e-05, "loss": 0.0691, "step": 70400 }, { "epoch": 7.09, "learning_rate": 3.2366084938969034e-05, "loss": 0.0662, "step": 70450 }, { "epoch": 7.09, "learning_rate": 3.23534752345405e-05, "loss": 0.0707, "step": 70500 }, { "epoch": 7.09, "eval_cer": 0.06729182622556804, "eval_loss": 0.4747462868690491, "eval_runtime": 1306.717, "eval_samples_per_second": 6.241, "eval_steps_per_second": 0.781, "step": 70500 }, { "epoch": 7.1, "learning_rate": 3.234086553011198e-05, "loss": 0.0702, "step": 70550 }, { "epoch": 7.1, "learning_rate": 3.2328255825683445e-05, "loss": 0.0712, "step": 70600 }, { "epoch": 7.11, "learning_rate": 3.231564612125492e-05, "loss": 0.0686, "step": 70650 }, { "epoch": 7.11, "learning_rate": 3.2303036416826395e-05, "loss": 0.0747, "step": 70700 }, { "epoch": 7.12, "learning_rate": 3.229042671239786e-05, "loss": 0.0698, "step": 70750 }, { "epoch": 7.12, "learning_rate": 3.227781700796934e-05, "loss": 0.0703, "step": 70800 }, { "epoch": 7.13, "learning_rate": 3.2265207303540806e-05, "loss": 0.0774, "step": 70850 }, { "epoch": 7.13, "learning_rate": 3.225259759911228e-05, "loss": 0.0722, "step": 70900 }, { "epoch": 7.14, "learning_rate": 3.223998789468375e-05, "loss": 0.0755, "step": 70950 }, { "epoch": 7.14, "learning_rate": 3.222737819025522e-05, "loss": 0.0696, "step": 71000 }, { "epoch": 7.14, "eval_cer": 0.06760627401166881, "eval_loss": 0.48029667139053345, "eval_runtime": 1308.9198, "eval_samples_per_second": 6.23, "eval_steps_per_second": 0.779, "step": 71000 }, { "epoch": 7.15, "learning_rate": 3.22147684858267e-05, "loss": 0.0711, "step": 71050 }, { "epoch": 7.15, "learning_rate": 3.2202158781398167e-05, "loss": 0.0725, "step": 71100 }, { "epoch": 7.16, "learning_rate": 3.218954907696964e-05, "loss": 0.0745, "step": 71150 }, { "epoch": 7.16, "learning_rate": 3.217693937254111e-05, "loss": 0.07, "step": 71200 }, { "epoch": 7.17, "learning_rate": 3.216432966811258e-05, "loss": 0.0676, "step": 71250 }, { "epoch": 7.17, "learning_rate": 3.215171996368405e-05, "loss": 0.0724, "step": 71300 }, { "epoch": 7.18, "learning_rate": 3.213911025925552e-05, "loss": 0.0794, "step": 71350 }, { "epoch": 7.18, "learning_rate": 3.2126500554826995e-05, "loss": 0.0781, "step": 71400 }, { "epoch": 7.19, "learning_rate": 3.211389085039847e-05, "loss": 0.0743, "step": 71450 }, { "epoch": 7.19, "learning_rate": 3.210128114596994e-05, "loss": 0.0735, "step": 71500 }, { "epoch": 7.19, "eval_cer": 0.06721399261514704, "eval_loss": 0.4748474955558777, "eval_runtime": 1326.8518, "eval_samples_per_second": 6.146, "eval_steps_per_second": 0.769, "step": 71500 }, { "epoch": 7.2, "learning_rate": 3.208867144154141e-05, "loss": 0.0776, "step": 71550 }, { "epoch": 7.2, "learning_rate": 3.207606173711288e-05, "loss": 0.0781, "step": 71600 }, { "epoch": 7.21, "learning_rate": 3.2063452032684356e-05, "loss": 0.0674, "step": 71650 }, { "epoch": 7.21, "learning_rate": 3.2050842328255824e-05, "loss": 0.0816, "step": 71700 }, { "epoch": 7.22, "learning_rate": 3.20382326238273e-05, "loss": 0.0737, "step": 71750 }, { "epoch": 7.22, "learning_rate": 3.2025622919398774e-05, "loss": 0.0713, "step": 71800 }, { "epoch": 7.23, "learning_rate": 3.201301321497024e-05, "loss": 0.07, "step": 71850 }, { "epoch": 7.23, "learning_rate": 3.200040351054172e-05, "loss": 0.0762, "step": 71900 }, { "epoch": 7.24, "learning_rate": 3.1987793806113185e-05, "loss": 0.078, "step": 71950 }, { "epoch": 7.24, "learning_rate": 3.197518410168466e-05, "loss": 0.0706, "step": 72000 }, { "epoch": 7.24, "eval_cer": 0.06685907135162734, "eval_loss": 0.47586891055107117, "eval_runtime": 1313.1843, "eval_samples_per_second": 6.21, "eval_steps_per_second": 0.777, "step": 72000 }, { "epoch": 7.25, "learning_rate": 3.196257439725613e-05, "loss": 0.0769, "step": 72050 }, { "epoch": 7.25, "learning_rate": 3.19499646928276e-05, "loss": 0.0892, "step": 72100 }, { "epoch": 7.26, "learning_rate": 3.193735498839907e-05, "loss": 0.0831, "step": 72150 }, { "epoch": 7.27, "learning_rate": 3.1924745283970546e-05, "loss": 0.0779, "step": 72200 }, { "epoch": 7.27, "learning_rate": 3.191213557954202e-05, "loss": 0.0806, "step": 72250 }, { "epoch": 7.28, "learning_rate": 3.189952587511349e-05, "loss": 0.0718, "step": 72300 }, { "epoch": 7.28, "learning_rate": 3.1886916170684964e-05, "loss": 0.0703, "step": 72350 }, { "epoch": 7.29, "learning_rate": 3.187430646625643e-05, "loss": 0.069, "step": 72400 }, { "epoch": 7.29, "learning_rate": 3.18616967618279e-05, "loss": 0.0724, "step": 72450 }, { "epoch": 7.3, "learning_rate": 3.1849087057399375e-05, "loss": 0.0802, "step": 72500 }, { "epoch": 7.3, "eval_cer": 0.0671392723491429, "eval_loss": 0.4817012846469879, "eval_runtime": 1310.121, "eval_samples_per_second": 6.225, "eval_steps_per_second": 0.779, "step": 72500 }, { "epoch": 7.3, "learning_rate": 3.183647735297085e-05, "loss": 0.0687, "step": 72550 }, { "epoch": 7.31, "learning_rate": 3.1823867648542325e-05, "loss": 0.0794, "step": 72600 }, { "epoch": 7.31, "learning_rate": 3.181125794411379e-05, "loss": 0.071, "step": 72650 }, { "epoch": 7.32, "learning_rate": 3.179864823968526e-05, "loss": 0.0815, "step": 72700 }, { "epoch": 7.32, "learning_rate": 3.1786038535256736e-05, "loss": 0.0809, "step": 72750 }, { "epoch": 7.33, "learning_rate": 3.1773428830828204e-05, "loss": 0.0758, "step": 72800 }, { "epoch": 7.33, "learning_rate": 3.176107132048825e-05, "loss": 0.0811, "step": 72850 }, { "epoch": 7.34, "learning_rate": 3.174846161605972e-05, "loss": 0.0738, "step": 72900 }, { "epoch": 7.34, "learning_rate": 3.173585191163119e-05, "loss": 0.0714, "step": 72950 }, { "epoch": 7.35, "learning_rate": 3.1723242207202666e-05, "loss": 0.0717, "step": 73000 }, { "epoch": 7.35, "eval_cer": 0.06699294516155144, "eval_loss": 0.47503411769866943, "eval_runtime": 1319.6258, "eval_samples_per_second": 6.18, "eval_steps_per_second": 0.773, "step": 73000 }, { "epoch": 7.35, "learning_rate": 3.1710632502774134e-05, "loss": 0.0871, "step": 73050 }, { "epoch": 7.36, "learning_rate": 3.16980227983456e-05, "loss": 0.0759, "step": 73100 }, { "epoch": 7.36, "learning_rate": 3.1685413093917084e-05, "loss": 0.0753, "step": 73150 }, { "epoch": 7.37, "learning_rate": 3.167280338948855e-05, "loss": 0.0827, "step": 73200 }, { "epoch": 7.37, "learning_rate": 3.166019368506003e-05, "loss": 0.0726, "step": 73250 }, { "epoch": 7.38, "learning_rate": 3.1647583980631495e-05, "loss": 0.0848, "step": 73300 }, { "epoch": 7.38, "learning_rate": 3.163497427620296e-05, "loss": 0.0758, "step": 73350 }, { "epoch": 7.39, "learning_rate": 3.162236457177444e-05, "loss": 0.0751, "step": 73400 }, { "epoch": 7.39, "learning_rate": 3.1609754867345906e-05, "loss": 0.0791, "step": 73450 }, { "epoch": 7.4, "learning_rate": 3.159714516291739e-05, "loss": 0.0735, "step": 73500 }, { "epoch": 7.4, "eval_cer": 0.06738522655807322, "eval_loss": 0.4762449264526367, "eval_runtime": 1314.4096, "eval_samples_per_second": 6.204, "eval_steps_per_second": 0.776, "step": 73500 }, { "epoch": 7.4, "learning_rate": 3.1584535458488856e-05, "loss": 0.0787, "step": 73550 }, { "epoch": 7.41, "learning_rate": 3.157192575406033e-05, "loss": 0.0813, "step": 73600 }, { "epoch": 7.41, "learning_rate": 3.15593160496318e-05, "loss": 0.081, "step": 73650 }, { "epoch": 7.42, "learning_rate": 3.154670634520327e-05, "loss": 0.0792, "step": 73700 }, { "epoch": 7.42, "learning_rate": 3.153409664077474e-05, "loss": 0.0814, "step": 73750 }, { "epoch": 7.43, "learning_rate": 3.152148693634621e-05, "loss": 0.0771, "step": 73800 }, { "epoch": 7.43, "learning_rate": 3.1508877231917685e-05, "loss": 0.0759, "step": 73850 }, { "epoch": 7.44, "learning_rate": 3.149626752748916e-05, "loss": 0.0877, "step": 73900 }, { "epoch": 7.44, "learning_rate": 3.148365782306063e-05, "loss": 0.0786, "step": 73950 }, { "epoch": 7.45, "learning_rate": 3.14710481186321e-05, "loss": 0.0685, "step": 74000 }, { "epoch": 7.45, "eval_cer": 0.06677189770795584, "eval_loss": 0.4741547405719757, "eval_runtime": 1319.1682, "eval_samples_per_second": 6.182, "eval_steps_per_second": 0.773, "step": 74000 }, { "epoch": 7.45, "learning_rate": 3.145843841420357e-05, "loss": 0.0746, "step": 74050 }, { "epoch": 7.46, "learning_rate": 3.1445828709775046e-05, "loss": 0.0764, "step": 74100 }, { "epoch": 7.46, "learning_rate": 3.1433219005346514e-05, "loss": 0.0789, "step": 74150 }, { "epoch": 7.47, "learning_rate": 3.142086149500656e-05, "loss": 0.0746, "step": 74200 }, { "epoch": 7.47, "learning_rate": 3.140825179057803e-05, "loss": 0.0791, "step": 74250 }, { "epoch": 7.48, "learning_rate": 3.13956420861495e-05, "loss": 0.0812, "step": 74300 }, { "epoch": 7.48, "learning_rate": 3.138303238172097e-05, "loss": 0.0813, "step": 74350 }, { "epoch": 7.49, "learning_rate": 3.1370422677292444e-05, "loss": 0.0919, "step": 74400 }, { "epoch": 7.49, "learning_rate": 3.135781297286392e-05, "loss": 0.0751, "step": 74450 }, { "epoch": 7.5, "learning_rate": 3.1345203268435394e-05, "loss": 0.0713, "step": 74500 }, { "epoch": 7.5, "eval_cer": 0.06723267268164808, "eval_loss": 0.47803500294685364, "eval_runtime": 1314.9256, "eval_samples_per_second": 6.202, "eval_steps_per_second": 0.776, "step": 74500 }, { "epoch": 7.5, "learning_rate": 3.133259356400686e-05, "loss": 0.0859, "step": 74550 }, { "epoch": 7.51, "learning_rate": 3.131998385957833e-05, "loss": 0.0788, "step": 74600 }, { "epoch": 7.51, "learning_rate": 3.1307374155149805e-05, "loss": 0.0933, "step": 74650 }, { "epoch": 7.52, "learning_rate": 3.129476445072127e-05, "loss": 0.0888, "step": 74700 }, { "epoch": 7.52, "learning_rate": 3.128215474629275e-05, "loss": 0.0769, "step": 74750 }, { "epoch": 7.53, "learning_rate": 3.1269545041864216e-05, "loss": 0.0805, "step": 74800 }, { "epoch": 7.53, "learning_rate": 3.125693533743569e-05, "loss": 0.0782, "step": 74850 }, { "epoch": 7.54, "learning_rate": 3.1244325633007166e-05, "loss": 0.0794, "step": 74900 }, { "epoch": 7.54, "learning_rate": 3.1231715928578634e-05, "loss": 0.0831, "step": 74950 }, { "epoch": 7.55, "learning_rate": 3.121910622415011e-05, "loss": 0.072, "step": 75000 }, { "epoch": 7.55, "eval_cer": 0.06595931481516074, "eval_loss": 0.47764837741851807, "eval_runtime": 1311.4531, "eval_samples_per_second": 6.218, "eval_steps_per_second": 0.778, "step": 75000 }, { "epoch": 7.55, "learning_rate": 3.120649651972158e-05, "loss": 0.0791, "step": 75050 }, { "epoch": 7.56, "learning_rate": 3.119388681529305e-05, "loss": 0.0766, "step": 75100 }, { "epoch": 7.56, "learning_rate": 3.118127711086452e-05, "loss": 0.0845, "step": 75150 }, { "epoch": 7.57, "learning_rate": 3.1168667406435995e-05, "loss": 0.073, "step": 75200 }, { "epoch": 7.57, "learning_rate": 3.115605770200747e-05, "loss": 0.0872, "step": 75250 }, { "epoch": 7.58, "learning_rate": 3.114344799757894e-05, "loss": 0.0864, "step": 75300 }, { "epoch": 7.58, "learning_rate": 3.113083829315041e-05, "loss": 0.0819, "step": 75350 }, { "epoch": 7.59, "learning_rate": 3.111822858872188e-05, "loss": 0.0809, "step": 75400 }, { "epoch": 7.59, "learning_rate": 3.1105618884293356e-05, "loss": 0.0837, "step": 75450 }, { "epoch": 7.6, "learning_rate": 3.1093009179864824e-05, "loss": 0.0841, "step": 75500 }, { "epoch": 7.6, "eval_cer": 0.06711747893822502, "eval_loss": 0.47860249876976013, "eval_runtime": 1315.1511, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 75500 }, { "epoch": 7.6, "learning_rate": 3.108039947543629e-05, "loss": 0.0875, "step": 75550 }, { "epoch": 7.61, "learning_rate": 3.1067789771007774e-05, "loss": 0.0731, "step": 75600 }, { "epoch": 7.61, "learning_rate": 3.105518006657924e-05, "loss": 0.0825, "step": 75650 }, { "epoch": 7.62, "learning_rate": 3.1042570362150717e-05, "loss": 0.0764, "step": 75700 }, { "epoch": 7.62, "learning_rate": 3.1029960657722185e-05, "loss": 0.0719, "step": 75750 }, { "epoch": 7.63, "learning_rate": 3.101735095329365e-05, "loss": 0.0823, "step": 75800 }, { "epoch": 7.63, "learning_rate": 3.100474124886513e-05, "loss": 0.0849, "step": 75850 }, { "epoch": 7.64, "learning_rate": 3.0992131544436596e-05, "loss": 0.0773, "step": 75900 }, { "epoch": 7.64, "learning_rate": 3.097952184000808e-05, "loss": 0.0849, "step": 75950 }, { "epoch": 7.65, "learning_rate": 3.0966912135579545e-05, "loss": 0.0933, "step": 76000 }, { "epoch": 7.65, "eval_cer": 0.06714238569355974, "eval_loss": 0.4770117402076721, "eval_runtime": 1378.743, "eval_samples_per_second": 5.915, "eval_steps_per_second": 0.74, "step": 76000 }, { "epoch": 7.65, "learning_rate": 3.0954302431151014e-05, "loss": 0.0861, "step": 76050 }, { "epoch": 7.66, "learning_rate": 3.094169272672249e-05, "loss": 0.0886, "step": 76100 }, { "epoch": 7.66, "learning_rate": 3.0929083022293957e-05, "loss": 0.1012, "step": 76150 }, { "epoch": 7.67, "learning_rate": 3.091647331786543e-05, "loss": 0.0815, "step": 76200 }, { "epoch": 7.67, "learning_rate": 3.09038636134369e-05, "loss": 0.0784, "step": 76250 }, { "epoch": 7.68, "learning_rate": 3.0891253909008374e-05, "loss": 0.0802, "step": 76300 }, { "epoch": 7.68, "learning_rate": 3.087864420457985e-05, "loss": 0.089, "step": 76350 }, { "epoch": 7.69, "learning_rate": 3.086603450015132e-05, "loss": 0.0839, "step": 76400 }, { "epoch": 7.69, "learning_rate": 3.085342479572279e-05, "loss": 0.0881, "step": 76450 }, { "epoch": 7.7, "learning_rate": 3.084081509129426e-05, "loss": 0.0734, "step": 76500 }, { "epoch": 7.7, "eval_cer": 0.06659755042061283, "eval_loss": 0.47463569045066833, "eval_runtime": 1460.8868, "eval_samples_per_second": 5.582, "eval_steps_per_second": 0.698, "step": 76500 }, { "epoch": 7.7, "learning_rate": 3.0828205386865735e-05, "loss": 0.078, "step": 76550 }, { "epoch": 7.71, "learning_rate": 3.08155956824372e-05, "loss": 0.0842, "step": 76600 }, { "epoch": 7.71, "learning_rate": 3.080298597800868e-05, "loss": 0.0794, "step": 76650 }, { "epoch": 7.72, "learning_rate": 3.0790376273580146e-05, "loss": 0.0869, "step": 76700 }, { "epoch": 7.72, "learning_rate": 3.077776656915162e-05, "loss": 0.0744, "step": 76750 }, { "epoch": 7.73, "learning_rate": 3.0765156864723096e-05, "loss": 0.0831, "step": 76800 }, { "epoch": 7.73, "learning_rate": 3.0752547160294564e-05, "loss": 0.0884, "step": 76850 }, { "epoch": 7.74, "learning_rate": 3.073993745586604e-05, "loss": 0.0813, "step": 76900 }, { "epoch": 7.74, "learning_rate": 3.072732775143751e-05, "loss": 0.0851, "step": 76950 }, { "epoch": 7.75, "learning_rate": 3.0714718047008975e-05, "loss": 0.0786, "step": 77000 }, { "epoch": 7.75, "eval_cer": 0.06698049178388409, "eval_loss": 0.4790664315223694, "eval_runtime": 1378.1687, "eval_samples_per_second": 5.917, "eval_steps_per_second": 0.74, "step": 77000 }, { "epoch": 7.75, "learning_rate": 3.070210834258045e-05, "loss": 0.0896, "step": 77050 }, { "epoch": 7.76, "learning_rate": 3.0689498638151925e-05, "loss": 0.0805, "step": 77100 }, { "epoch": 7.76, "learning_rate": 3.06768889337234e-05, "loss": 0.082, "step": 77150 }, { "epoch": 7.77, "learning_rate": 3.066427922929487e-05, "loss": 0.0845, "step": 77200 }, { "epoch": 7.77, "learning_rate": 3.0651669524866336e-05, "loss": 0.08, "step": 77250 }, { "epoch": 7.78, "learning_rate": 3.063905982043781e-05, "loss": 0.0804, "step": 77300 }, { "epoch": 7.78, "learning_rate": 3.062645011600928e-05, "loss": 0.0918, "step": 77350 }, { "epoch": 7.79, "learning_rate": 3.0613840411580754e-05, "loss": 0.0806, "step": 77400 }, { "epoch": 7.79, "learning_rate": 3.060123070715222e-05, "loss": 0.08, "step": 77450 }, { "epoch": 7.8, "learning_rate": 3.05886210027237e-05, "loss": 0.087, "step": 77500 }, { "epoch": 7.8, "eval_cer": 0.06695869837296621, "eval_loss": 0.47620683908462524, "eval_runtime": 1348.0373, "eval_samples_per_second": 6.05, "eval_steps_per_second": 0.757, "step": 77500 }, { "epoch": 7.8, "learning_rate": 3.057626349238374e-05, "loss": 0.0744, "step": 77550 }, { "epoch": 7.81, "learning_rate": 3.056365378795521e-05, "loss": 0.0896, "step": 77600 }, { "epoch": 7.81, "learning_rate": 3.055104408352668e-05, "loss": 0.0823, "step": 77650 }, { "epoch": 7.82, "learning_rate": 3.053843437909816e-05, "loss": 0.082, "step": 77700 }, { "epoch": 7.82, "learning_rate": 3.052582467466963e-05, "loss": 0.0822, "step": 77750 }, { "epoch": 7.83, "learning_rate": 3.05132149702411e-05, "loss": 0.0917, "step": 77800 }, { "epoch": 7.83, "learning_rate": 3.050060526581257e-05, "loss": 0.0848, "step": 77850 }, { "epoch": 7.84, "learning_rate": 3.0487995561384042e-05, "loss": 0.0899, "step": 77900 }, { "epoch": 7.84, "learning_rate": 3.0475385856955513e-05, "loss": 0.078, "step": 77950 }, { "epoch": 7.85, "learning_rate": 3.0462776152526985e-05, "loss": 0.0865, "step": 78000 }, { "epoch": 7.85, "eval_cer": 0.06584100772732085, "eval_loss": 0.4775438606739044, "eval_runtime": 1339.2592, "eval_samples_per_second": 6.089, "eval_steps_per_second": 0.762, "step": 78000 }, { "epoch": 7.85, "learning_rate": 3.045016644809846e-05, "loss": 0.0881, "step": 78050 }, { "epoch": 7.86, "learning_rate": 3.043755674366993e-05, "loss": 0.0857, "step": 78100 }, { "epoch": 7.86, "learning_rate": 3.0424947039241403e-05, "loss": 0.0922, "step": 78150 }, { "epoch": 7.87, "learning_rate": 3.0412337334812874e-05, "loss": 0.0777, "step": 78200 }, { "epoch": 7.87, "learning_rate": 3.0399727630384346e-05, "loss": 0.0768, "step": 78250 }, { "epoch": 7.88, "learning_rate": 3.0387117925955817e-05, "loss": 0.09, "step": 78300 }, { "epoch": 7.88, "learning_rate": 3.037450822152729e-05, "loss": 0.0769, "step": 78350 }, { "epoch": 7.89, "learning_rate": 3.0361898517098757e-05, "loss": 0.0767, "step": 78400 }, { "epoch": 7.89, "learning_rate": 3.0349288812670235e-05, "loss": 0.0952, "step": 78450 }, { "epoch": 7.9, "learning_rate": 3.0336931302330273e-05, "loss": 0.088, "step": 78500 }, { "epoch": 7.9, "eval_cer": 0.06645433657743821, "eval_loss": 0.4749051034450531, "eval_runtime": 1304.1372, "eval_samples_per_second": 6.253, "eval_steps_per_second": 0.782, "step": 78500 }, { "epoch": 7.9, "learning_rate": 3.0324321597901744e-05, "loss": 0.0785, "step": 78550 }, { "epoch": 7.91, "learning_rate": 3.0311711893473216e-05, "loss": 0.0845, "step": 78600 }, { "epoch": 7.91, "learning_rate": 3.0299102189044694e-05, "loss": 0.083, "step": 78650 }, { "epoch": 7.92, "learning_rate": 3.0286492484616162e-05, "loss": 0.0831, "step": 78700 }, { "epoch": 7.92, "learning_rate": 3.0273882780187634e-05, "loss": 0.0897, "step": 78750 }, { "epoch": 7.93, "learning_rate": 3.0261273075759105e-05, "loss": 0.0788, "step": 78800 }, { "epoch": 7.93, "learning_rate": 3.0248663371330577e-05, "loss": 0.0932, "step": 78850 }, { "epoch": 7.94, "learning_rate": 3.0236053666902048e-05, "loss": 0.0808, "step": 78900 }, { "epoch": 7.94, "learning_rate": 3.022344396247352e-05, "loss": 0.0847, "step": 78950 }, { "epoch": 7.95, "learning_rate": 3.0210834258044994e-05, "loss": 0.0861, "step": 79000 }, { "epoch": 7.95, "eval_cer": 0.0655141065635527, "eval_loss": 0.4719955325126648, "eval_runtime": 1326.488, "eval_samples_per_second": 6.148, "eval_steps_per_second": 0.769, "step": 79000 }, { "epoch": 7.95, "learning_rate": 3.0198224553616466e-05, "loss": 0.0781, "step": 79050 }, { "epoch": 7.96, "learning_rate": 3.0185614849187937e-05, "loss": 0.086, "step": 79100 }, { "epoch": 7.96, "learning_rate": 3.017300514475941e-05, "loss": 0.0822, "step": 79150 }, { "epoch": 7.97, "learning_rate": 3.016039544033088e-05, "loss": 0.0861, "step": 79200 }, { "epoch": 7.97, "learning_rate": 3.0147785735902352e-05, "loss": 0.0826, "step": 79250 }, { "epoch": 7.98, "learning_rate": 3.0135176031473823e-05, "loss": 0.0832, "step": 79300 }, { "epoch": 7.98, "learning_rate": 3.012256632704529e-05, "loss": 0.0865, "step": 79350 }, { "epoch": 7.99, "learning_rate": 3.010995662261677e-05, "loss": 0.084, "step": 79400 }, { "epoch": 7.99, "learning_rate": 3.009734691818824e-05, "loss": 0.0805, "step": 79450 }, { "epoch": 8.0, "learning_rate": 3.0084737213759713e-05, "loss": 0.0882, "step": 79500 }, { "epoch": 8.0, "eval_cer": 0.06565420706231047, "eval_loss": 0.47265464067459106, "eval_runtime": 1315.7561, "eval_samples_per_second": 6.198, "eval_steps_per_second": 0.775, "step": 79500 }, { "epoch": 8.0, "learning_rate": 3.0072127509331184e-05, "loss": 0.0547, "step": 79550 }, { "epoch": 8.01, "learning_rate": 3.0059517804902652e-05, "loss": 0.0527, "step": 79600 }, { "epoch": 8.01, "learning_rate": 3.0046908100474124e-05, "loss": 0.0538, "step": 79650 }, { "epoch": 8.02, "learning_rate": 3.0034298396045595e-05, "loss": 0.049, "step": 79700 }, { "epoch": 8.02, "learning_rate": 3.0021688691617073e-05, "loss": 0.054, "step": 79750 }, { "epoch": 8.03, "learning_rate": 3.0009078987188545e-05, "loss": 0.0535, "step": 79800 }, { "epoch": 8.03, "learning_rate": 2.9996469282760013e-05, "loss": 0.0523, "step": 79850 }, { "epoch": 8.04, "learning_rate": 2.9983859578331485e-05, "loss": 0.0462, "step": 79900 }, { "epoch": 8.04, "learning_rate": 2.9971249873902956e-05, "loss": 0.0548, "step": 79950 }, { "epoch": 8.05, "learning_rate": 2.9958640169474427e-05, "loss": 0.0496, "step": 80000 }, { "epoch": 8.05, "eval_cer": 0.06602780839233122, "eval_loss": 0.4890212416648865, "eval_runtime": 1634.2805, "eval_samples_per_second": 4.99, "eval_steps_per_second": 0.624, "step": 80000 }, { "epoch": 8.05, "learning_rate": 2.99460304650459e-05, "loss": 0.0487, "step": 80050 }, { "epoch": 8.06, "learning_rate": 2.993342076061737e-05, "loss": 0.0495, "step": 80100 }, { "epoch": 8.06, "learning_rate": 2.9920811056188845e-05, "loss": 0.0674, "step": 80150 }, { "epoch": 8.07, "learning_rate": 2.9908201351760317e-05, "loss": 0.0484, "step": 80200 }, { "epoch": 8.08, "learning_rate": 2.9895591647331788e-05, "loss": 0.0489, "step": 80250 }, { "epoch": 8.08, "learning_rate": 2.988298194290326e-05, "loss": 0.0499, "step": 80300 }, { "epoch": 8.09, "learning_rate": 2.987037223847473e-05, "loss": 0.0582, "step": 80350 }, { "epoch": 8.09, "learning_rate": 2.9857762534046203e-05, "loss": 0.0528, "step": 80400 }, { "epoch": 8.1, "learning_rate": 2.9845152829617674e-05, "loss": 0.0598, "step": 80450 }, { "epoch": 8.1, "learning_rate": 2.983254312518915e-05, "loss": 0.0484, "step": 80500 }, { "epoch": 8.1, "eval_cer": 0.0649381378464374, "eval_loss": 0.4887416958808899, "eval_runtime": 1631.311, "eval_samples_per_second": 4.999, "eval_steps_per_second": 0.625, "step": 80500 }, { "epoch": 8.11, "learning_rate": 2.981993342076062e-05, "loss": 0.0532, "step": 80550 }, { "epoch": 8.11, "learning_rate": 2.9807323716332092e-05, "loss": 0.0535, "step": 80600 }, { "epoch": 8.12, "learning_rate": 2.979496620599213e-05, "loss": 0.0611, "step": 80650 }, { "epoch": 8.12, "learning_rate": 2.9782356501563608e-05, "loss": 0.055, "step": 80700 }, { "epoch": 8.13, "learning_rate": 2.976974679713508e-05, "loss": 0.0539, "step": 80750 }, { "epoch": 8.13, "learning_rate": 2.975713709270655e-05, "loss": 0.0495, "step": 80800 }, { "epoch": 8.14, "learning_rate": 2.974452738827802e-05, "loss": 0.0583, "step": 80850 }, { "epoch": 8.14, "learning_rate": 2.973191768384949e-05, "loss": 0.0597, "step": 80900 }, { "epoch": 8.15, "learning_rate": 2.9719307979420962e-05, "loss": 0.0484, "step": 80950 }, { "epoch": 8.15, "learning_rate": 2.9706698274992434e-05, "loss": 0.0584, "step": 81000 }, { "epoch": 8.15, "eval_cer": 0.06503776486777627, "eval_loss": 0.48485833406448364, "eval_runtime": 1640.6666, "eval_samples_per_second": 4.971, "eval_steps_per_second": 0.622, "step": 81000 }, { "epoch": 8.16, "learning_rate": 2.9694088570563905e-05, "loss": 0.0559, "step": 81050 }, { "epoch": 8.16, "learning_rate": 2.968147886613538e-05, "loss": 0.0502, "step": 81100 }, { "epoch": 8.17, "learning_rate": 2.966886916170685e-05, "loss": 0.0541, "step": 81150 }, { "epoch": 8.17, "learning_rate": 2.9656259457278323e-05, "loss": 0.0484, "step": 81200 }, { "epoch": 8.18, "learning_rate": 2.9643649752849795e-05, "loss": 0.0542, "step": 81250 }, { "epoch": 8.18, "learning_rate": 2.9631040048421266e-05, "loss": 0.0505, "step": 81300 }, { "epoch": 8.19, "learning_rate": 2.9618430343992738e-05, "loss": 0.0502, "step": 81350 }, { "epoch": 8.19, "learning_rate": 2.960582063956421e-05, "loss": 0.0554, "step": 81400 }, { "epoch": 8.2, "learning_rate": 2.9593210935135684e-05, "loss": 0.0526, "step": 81450 }, { "epoch": 8.2, "learning_rate": 2.9580601230707155e-05, "loss": 0.0628, "step": 81500 }, { "epoch": 8.2, "eval_cer": 0.06687775141812838, "eval_loss": 0.48889070749282837, "eval_runtime": 1639.6091, "eval_samples_per_second": 4.974, "eval_steps_per_second": 0.622, "step": 81500 }, { "epoch": 8.21, "learning_rate": 2.9567991526278627e-05, "loss": 0.0522, "step": 81550 }, { "epoch": 8.21, "learning_rate": 2.95553818218501e-05, "loss": 0.0534, "step": 81600 }, { "epoch": 8.22, "learning_rate": 2.954277211742157e-05, "loss": 0.052, "step": 81650 }, { "epoch": 8.22, "learning_rate": 2.953016241299304e-05, "loss": 0.0562, "step": 81700 }, { "epoch": 8.23, "learning_rate": 2.951755270856451e-05, "loss": 0.0532, "step": 81750 }, { "epoch": 8.23, "learning_rate": 2.950494300413598e-05, "loss": 0.0549, "step": 81800 }, { "epoch": 8.24, "learning_rate": 2.949233329970746e-05, "loss": 0.0569, "step": 81850 }, { "epoch": 8.24, "learning_rate": 2.947972359527893e-05, "loss": 0.0531, "step": 81900 }, { "epoch": 8.25, "learning_rate": 2.9467113890850402e-05, "loss": 0.0583, "step": 81950 }, { "epoch": 8.25, "learning_rate": 2.945450418642187e-05, "loss": 0.0535, "step": 82000 }, { "epoch": 8.25, "eval_cer": 0.065411366197797, "eval_loss": 0.48994818329811096, "eval_runtime": 1667.4508, "eval_samples_per_second": 4.891, "eval_steps_per_second": 0.612, "step": 82000 }, { "epoch": 8.26, "learning_rate": 2.9441894481993342e-05, "loss": 0.0554, "step": 82050 }, { "epoch": 8.26, "learning_rate": 2.9429284777564813e-05, "loss": 0.0602, "step": 82100 }, { "epoch": 8.27, "learning_rate": 2.9416675073136285e-05, "loss": 0.0518, "step": 82150 }, { "epoch": 8.27, "learning_rate": 2.9404065368707763e-05, "loss": 0.0555, "step": 82200 }, { "epoch": 8.28, "learning_rate": 2.939145566427923e-05, "loss": 0.0515, "step": 82250 }, { "epoch": 8.28, "learning_rate": 2.9378845959850703e-05, "loss": 0.0567, "step": 82300 }, { "epoch": 8.29, "learning_rate": 2.9366236255422174e-05, "loss": 0.0624, "step": 82350 }, { "epoch": 8.29, "learning_rate": 2.9353626550993645e-05, "loss": 0.0571, "step": 82400 }, { "epoch": 8.3, "learning_rate": 2.9341016846565117e-05, "loss": 0.0561, "step": 82450 }, { "epoch": 8.3, "learning_rate": 2.932840714213659e-05, "loss": 0.0603, "step": 82500 }, { "epoch": 8.3, "eval_cer": 0.06593752140424286, "eval_loss": 0.48891210556030273, "eval_runtime": 1669.5503, "eval_samples_per_second": 4.885, "eval_steps_per_second": 0.611, "step": 82500 }, { "epoch": 8.31, "learning_rate": 2.931579743770806e-05, "loss": 0.0605, "step": 82550 }, { "epoch": 8.31, "learning_rate": 2.9303187733279535e-05, "loss": 0.0525, "step": 82600 }, { "epoch": 8.32, "learning_rate": 2.9290578028851006e-05, "loss": 0.0657, "step": 82650 }, { "epoch": 8.32, "learning_rate": 2.9277968324422478e-05, "loss": 0.0591, "step": 82700 }, { "epoch": 8.33, "learning_rate": 2.926535861999395e-05, "loss": 0.0598, "step": 82750 }, { "epoch": 8.33, "learning_rate": 2.925274891556542e-05, "loss": 0.0506, "step": 82800 }, { "epoch": 8.34, "learning_rate": 2.9240139211136892e-05, "loss": 0.051, "step": 82850 }, { "epoch": 8.34, "learning_rate": 2.9227529506708364e-05, "loss": 0.0508, "step": 82900 }, { "epoch": 8.35, "learning_rate": 2.921491980227984e-05, "loss": 0.0609, "step": 82950 }, { "epoch": 8.35, "learning_rate": 2.920231009785131e-05, "loss": 0.055, "step": 83000 }, { "epoch": 8.35, "eval_cer": 0.06559816686280737, "eval_loss": 0.489936888217926, "eval_runtime": 1658.8433, "eval_samples_per_second": 4.916, "eval_steps_per_second": 0.615, "step": 83000 }, { "epoch": 8.36, "learning_rate": 2.918970039342278e-05, "loss": 0.0583, "step": 83050 }, { "epoch": 8.36, "learning_rate": 2.9177090688994253e-05, "loss": 0.054, "step": 83100 }, { "epoch": 8.37, "learning_rate": 2.9164480984565724e-05, "loss": 0.0571, "step": 83150 }, { "epoch": 8.37, "learning_rate": 2.9151871280137193e-05, "loss": 0.0506, "step": 83200 }, { "epoch": 8.38, "learning_rate": 2.9139261575708664e-05, "loss": 0.0569, "step": 83250 }, { "epoch": 8.38, "learning_rate": 2.9126651871280136e-05, "loss": 0.0555, "step": 83300 }, { "epoch": 8.39, "learning_rate": 2.9114042166851614e-05, "loss": 0.0619, "step": 83350 }, { "epoch": 8.39, "learning_rate": 2.9101432462423085e-05, "loss": 0.0611, "step": 83400 }, { "epoch": 8.4, "learning_rate": 2.9088822757994553e-05, "loss": 0.0609, "step": 83450 }, { "epoch": 8.4, "learning_rate": 2.9076213053566025e-05, "loss": 0.0587, "step": 83500 }, { "epoch": 8.4, "eval_cer": 0.0657662874613167, "eval_loss": 0.4944039583206177, "eval_runtime": 1594.4886, "eval_samples_per_second": 5.114, "eval_steps_per_second": 0.64, "step": 83500 }, { "epoch": 8.41, "learning_rate": 2.9063603349137496e-05, "loss": 0.0511, "step": 83550 }, { "epoch": 8.41, "learning_rate": 2.9050993644708968e-05, "loss": 0.0642, "step": 83600 }, { "epoch": 8.42, "learning_rate": 2.903838394028044e-05, "loss": 0.0685, "step": 83650 }, { "epoch": 8.42, "learning_rate": 2.9025774235851914e-05, "loss": 0.0569, "step": 83700 }, { "epoch": 8.43, "learning_rate": 2.9013164531423386e-05, "loss": 0.0603, "step": 83750 }, { "epoch": 8.43, "learning_rate": 2.9000554826994857e-05, "loss": 0.0598, "step": 83800 }, { "epoch": 8.44, "learning_rate": 2.898794512256633e-05, "loss": 0.0682, "step": 83850 }, { "epoch": 8.44, "learning_rate": 2.89753354181378e-05, "loss": 0.0559, "step": 83900 }, { "epoch": 8.45, "learning_rate": 2.896272571370927e-05, "loss": 0.063, "step": 83950 }, { "epoch": 8.45, "learning_rate": 2.8950116009280743e-05, "loss": 0.057, "step": 84000 }, { "epoch": 8.45, "eval_cer": 0.06579742090548509, "eval_loss": 0.4896000623703003, "eval_runtime": 1368.5588, "eval_samples_per_second": 5.959, "eval_steps_per_second": 0.745, "step": 84000 }, { "epoch": 8.46, "learning_rate": 2.8937506304852215e-05, "loss": 0.0598, "step": 84050 }, { "epoch": 8.46, "learning_rate": 2.892489660042369e-05, "loss": 0.0574, "step": 84100 }, { "epoch": 8.47, "learning_rate": 2.891228689599516e-05, "loss": 0.0714, "step": 84150 }, { "epoch": 8.47, "learning_rate": 2.8899677191566632e-05, "loss": 0.0578, "step": 84200 }, { "epoch": 8.48, "learning_rate": 2.8887067487138104e-05, "loss": 0.0582, "step": 84250 }, { "epoch": 8.48, "learning_rate": 2.8874457782709575e-05, "loss": 0.0633, "step": 84300 }, { "epoch": 8.49, "learning_rate": 2.8861848078281047e-05, "loss": 0.0583, "step": 84350 }, { "epoch": 8.49, "learning_rate": 2.8849238373852515e-05, "loss": 0.0615, "step": 84400 }, { "epoch": 8.5, "learning_rate": 2.8836628669423986e-05, "loss": 0.0561, "step": 84450 }, { "epoch": 8.5, "learning_rate": 2.8824018964995465e-05, "loss": 0.0522, "step": 84500 }, { "epoch": 8.5, "eval_cer": 0.06557014676305581, "eval_loss": 0.4841054379940033, "eval_runtime": 1310.2215, "eval_samples_per_second": 6.224, "eval_steps_per_second": 0.778, "step": 84500 }, { "epoch": 8.51, "learning_rate": 2.8811409260566936e-05, "loss": 0.0591, "step": 84550 }, { "epoch": 8.51, "learning_rate": 2.8798799556138408e-05, "loss": 0.0614, "step": 84600 }, { "epoch": 8.52, "learning_rate": 2.8786189851709876e-05, "loss": 0.0601, "step": 84650 }, { "epoch": 8.52, "learning_rate": 2.8773580147281347e-05, "loss": 0.0561, "step": 84700 }, { "epoch": 8.53, "learning_rate": 2.876097044285282e-05, "loss": 0.0556, "step": 84750 }, { "epoch": 8.53, "learning_rate": 2.874836073842429e-05, "loss": 0.0581, "step": 84800 }, { "epoch": 8.54, "learning_rate": 2.873575103399577e-05, "loss": 0.0548, "step": 84850 }, { "epoch": 8.54, "learning_rate": 2.8723141329567237e-05, "loss": 0.0584, "step": 84900 }, { "epoch": 8.55, "learning_rate": 2.8710531625138708e-05, "loss": 0.0592, "step": 84950 }, { "epoch": 8.55, "learning_rate": 2.869792192071018e-05, "loss": 0.0627, "step": 85000 }, { "epoch": 8.55, "eval_cer": 0.06540825285338016, "eval_loss": 0.4865466058254242, "eval_runtime": 1318.1722, "eval_samples_per_second": 6.187, "eval_steps_per_second": 0.774, "step": 85000 }, { "epoch": 8.56, "learning_rate": 2.868531221628165e-05, "loss": 0.0588, "step": 85050 }, { "epoch": 8.56, "learning_rate": 2.8672702511853123e-05, "loss": 0.0571, "step": 85100 }, { "epoch": 8.57, "learning_rate": 2.8660092807424594e-05, "loss": 0.0639, "step": 85150 }, { "epoch": 8.57, "learning_rate": 2.8647483102996065e-05, "loss": 0.0597, "step": 85200 }, { "epoch": 8.58, "learning_rate": 2.863512559265611e-05, "loss": 0.0607, "step": 85250 }, { "epoch": 8.58, "learning_rate": 2.862251588822758e-05, "loss": 0.0588, "step": 85300 }, { "epoch": 8.59, "learning_rate": 2.860990618379905e-05, "loss": 0.0534, "step": 85350 }, { "epoch": 8.59, "learning_rate": 2.859729647937052e-05, "loss": 0.0533, "step": 85400 }, { "epoch": 8.6, "learning_rate": 2.8584686774942e-05, "loss": 0.0615, "step": 85450 }, { "epoch": 8.6, "learning_rate": 2.857207707051347e-05, "loss": 0.066, "step": 85500 }, { "epoch": 8.6, "eval_cer": 0.06566977378439467, "eval_loss": 0.48636212944984436, "eval_runtime": 1461.8658, "eval_samples_per_second": 5.578, "eval_steps_per_second": 0.698, "step": 85500 }, { "epoch": 8.61, "learning_rate": 2.8559467366084942e-05, "loss": 0.0609, "step": 85550 }, { "epoch": 8.61, "learning_rate": 2.854685766165641e-05, "loss": 0.0654, "step": 85600 }, { "epoch": 8.62, "learning_rate": 2.8534247957227882e-05, "loss": 0.0567, "step": 85650 }, { "epoch": 8.62, "learning_rate": 2.8521638252799354e-05, "loss": 0.0616, "step": 85700 }, { "epoch": 8.63, "learning_rate": 2.8509028548370825e-05, "loss": 0.0546, "step": 85750 }, { "epoch": 8.63, "learning_rate": 2.8496418843942303e-05, "loss": 0.0607, "step": 85800 }, { "epoch": 8.64, "learning_rate": 2.848380913951377e-05, "loss": 0.0621, "step": 85850 }, { "epoch": 8.64, "learning_rate": 2.8471199435085243e-05, "loss": 0.0562, "step": 85900 }, { "epoch": 8.65, "learning_rate": 2.8458589730656714e-05, "loss": 0.0679, "step": 85950 }, { "epoch": 8.65, "learning_rate": 2.8445980026228186e-05, "loss": 0.0575, "step": 86000 }, { "epoch": 8.65, "eval_cer": 0.06554524000772109, "eval_loss": 0.48721444606781006, "eval_runtime": 1622.6638, "eval_samples_per_second": 5.026, "eval_steps_per_second": 0.629, "step": 86000 }, { "epoch": 8.66, "learning_rate": 2.8433370321799657e-05, "loss": 0.0698, "step": 86050 }, { "epoch": 8.66, "learning_rate": 2.842076061737113e-05, "loss": 0.0603, "step": 86100 }, { "epoch": 8.67, "learning_rate": 2.84081509129426e-05, "loss": 0.0551, "step": 86150 }, { "epoch": 8.67, "learning_rate": 2.8395541208514075e-05, "loss": 0.054, "step": 86200 }, { "epoch": 8.68, "learning_rate": 2.8382931504085547e-05, "loss": 0.0573, "step": 86250 }, { "epoch": 8.68, "learning_rate": 2.8370321799657018e-05, "loss": 0.0644, "step": 86300 }, { "epoch": 8.69, "learning_rate": 2.835771209522849e-05, "loss": 0.058, "step": 86350 }, { "epoch": 8.69, "learning_rate": 2.8345354584888534e-05, "loss": 0.0612, "step": 86400 }, { "epoch": 8.7, "learning_rate": 2.8332744880460006e-05, "loss": 0.0644, "step": 86450 }, { "epoch": 8.7, "learning_rate": 2.8320135176031477e-05, "loss": 0.0691, "step": 86500 }, { "epoch": 8.7, "eval_cer": 0.0664138631000193, "eval_loss": 0.4914659261703491, "eval_runtime": 1367.2135, "eval_samples_per_second": 5.965, "eval_steps_per_second": 0.746, "step": 86500 }, { "epoch": 8.71, "learning_rate": 2.8307525471602945e-05, "loss": 0.0599, "step": 86550 }, { "epoch": 8.71, "learning_rate": 2.8294915767174417e-05, "loss": 0.0583, "step": 86600 }, { "epoch": 8.72, "learning_rate": 2.828230606274589e-05, "loss": 0.0601, "step": 86650 }, { "epoch": 8.72, "learning_rate": 2.826969635831736e-05, "loss": 0.0569, "step": 86700 }, { "epoch": 8.73, "learning_rate": 2.8257086653888838e-05, "loss": 0.0617, "step": 86750 }, { "epoch": 8.73, "learning_rate": 2.824447694946031e-05, "loss": 0.059, "step": 86800 }, { "epoch": 8.74, "learning_rate": 2.8231867245031778e-05, "loss": 0.0566, "step": 86850 }, { "epoch": 8.74, "learning_rate": 2.821925754060325e-05, "loss": 0.0593, "step": 86900 }, { "epoch": 8.75, "learning_rate": 2.820664783617472e-05, "loss": 0.0639, "step": 86950 }, { "epoch": 8.75, "learning_rate": 2.8194038131746192e-05, "loss": 0.0645, "step": 87000 }, { "epoch": 8.75, "eval_cer": 0.06538957278687912, "eval_loss": 0.48836806416511536, "eval_runtime": 1317.3561, "eval_samples_per_second": 6.19, "eval_steps_per_second": 0.774, "step": 87000 }, { "epoch": 8.76, "learning_rate": 2.8181428427317664e-05, "loss": 0.059, "step": 87050 }, { "epoch": 8.76, "learning_rate": 2.8168818722889135e-05, "loss": 0.0574, "step": 87100 }, { "epoch": 8.77, "learning_rate": 2.815620901846061e-05, "loss": 0.0608, "step": 87150 }, { "epoch": 8.77, "learning_rate": 2.814359931403208e-05, "loss": 0.0563, "step": 87200 }, { "epoch": 8.78, "learning_rate": 2.8130989609603553e-05, "loss": 0.0648, "step": 87250 }, { "epoch": 8.78, "learning_rate": 2.8118379905175024e-05, "loss": 0.0626, "step": 87300 }, { "epoch": 8.79, "learning_rate": 2.8105770200746496e-05, "loss": 0.0593, "step": 87350 }, { "epoch": 8.79, "learning_rate": 2.8093160496317967e-05, "loss": 0.0618, "step": 87400 }, { "epoch": 8.8, "learning_rate": 2.808055079188944e-05, "loss": 0.0588, "step": 87450 }, { "epoch": 8.8, "learning_rate": 2.8067941087460914e-05, "loss": 0.0644, "step": 87500 }, { "epoch": 8.8, "eval_cer": 0.0662395158126763, "eval_loss": 0.48550212383270264, "eval_runtime": 1317.7781, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 87500 }, { "epoch": 8.81, "learning_rate": 2.8055331383032385e-05, "loss": 0.0674, "step": 87550 }, { "epoch": 8.81, "learning_rate": 2.8042721678603857e-05, "loss": 0.065, "step": 87600 }, { "epoch": 8.82, "learning_rate": 2.8030111974175328e-05, "loss": 0.0578, "step": 87650 }, { "epoch": 8.82, "learning_rate": 2.80175022697468e-05, "loss": 0.0648, "step": 87700 }, { "epoch": 8.83, "learning_rate": 2.8004892565318268e-05, "loss": 0.0728, "step": 87750 }, { "epoch": 8.83, "learning_rate": 2.799228286088974e-05, "loss": 0.0626, "step": 87800 }, { "epoch": 8.84, "learning_rate": 2.797967315646121e-05, "loss": 0.0663, "step": 87850 }, { "epoch": 8.84, "learning_rate": 2.796706345203269e-05, "loss": 0.0625, "step": 87900 }, { "epoch": 8.85, "learning_rate": 2.795445374760416e-05, "loss": 0.059, "step": 87950 }, { "epoch": 8.85, "learning_rate": 2.794184404317563e-05, "loss": 0.0554, "step": 88000 }, { "epoch": 8.85, "eval_cer": 0.06522767887720347, "eval_loss": 0.4852359890937805, "eval_runtime": 1314.5728, "eval_samples_per_second": 6.204, "eval_steps_per_second": 0.776, "step": 88000 }, { "epoch": 8.86, "learning_rate": 2.79292343387471e-05, "loss": 0.0618, "step": 88050 }, { "epoch": 8.86, "learning_rate": 2.791662463431857e-05, "loss": 0.056, "step": 88100 }, { "epoch": 8.87, "learning_rate": 2.7904014929890043e-05, "loss": 0.0588, "step": 88150 }, { "epoch": 8.87, "learning_rate": 2.7891405225461514e-05, "loss": 0.0598, "step": 88200 }, { "epoch": 8.88, "learning_rate": 2.787879552103299e-05, "loss": 0.0591, "step": 88250 }, { "epoch": 8.89, "learning_rate": 2.786618581660446e-05, "loss": 0.0629, "step": 88300 }, { "epoch": 8.89, "learning_rate": 2.7853576112175932e-05, "loss": 0.0614, "step": 88350 }, { "epoch": 8.9, "learning_rate": 2.7840966407747404e-05, "loss": 0.0608, "step": 88400 }, { "epoch": 8.9, "learning_rate": 2.7828356703318875e-05, "loss": 0.0618, "step": 88450 }, { "epoch": 8.91, "learning_rate": 2.7815746998890347e-05, "loss": 0.0682, "step": 88500 }, { "epoch": 8.91, "eval_cer": 0.06553589997447058, "eval_loss": 0.4838888645172119, "eval_runtime": 1326.4968, "eval_samples_per_second": 6.148, "eval_steps_per_second": 0.769, "step": 88500 }, { "epoch": 8.91, "learning_rate": 2.7803137294461818e-05, "loss": 0.0627, "step": 88550 }, { "epoch": 8.92, "learning_rate": 2.779052759003329e-05, "loss": 0.0634, "step": 88600 }, { "epoch": 8.92, "learning_rate": 2.7777917885604765e-05, "loss": 0.0589, "step": 88650 }, { "epoch": 8.93, "learning_rate": 2.7765308181176236e-05, "loss": 0.0652, "step": 88700 }, { "epoch": 8.93, "learning_rate": 2.7752698476747708e-05, "loss": 0.0616, "step": 88750 }, { "epoch": 8.94, "learning_rate": 2.774008877231918e-05, "loss": 0.0644, "step": 88800 }, { "epoch": 8.94, "learning_rate": 2.772747906789065e-05, "loss": 0.0612, "step": 88850 }, { "epoch": 8.95, "learning_rate": 2.7714869363462122e-05, "loss": 0.0592, "step": 88900 }, { "epoch": 8.95, "learning_rate": 2.770225965903359e-05, "loss": 0.063, "step": 88950 }, { "epoch": 8.96, "learning_rate": 2.768964995460507e-05, "loss": 0.0725, "step": 89000 }, { "epoch": 8.96, "eval_cer": 0.06495370456852159, "eval_loss": 0.47889894247055054, "eval_runtime": 1411.279, "eval_samples_per_second": 5.778, "eval_steps_per_second": 0.723, "step": 89000 }, { "epoch": 8.96, "learning_rate": 2.767704025017654e-05, "loss": 0.0663, "step": 89050 }, { "epoch": 8.97, "learning_rate": 2.766443054574801e-05, "loss": 0.0648, "step": 89100 }, { "epoch": 8.97, "learning_rate": 2.7651820841319483e-05, "loss": 0.0655, "step": 89150 }, { "epoch": 8.98, "learning_rate": 2.763921113689095e-05, "loss": 0.067, "step": 89200 }, { "epoch": 8.98, "learning_rate": 2.7626601432462422e-05, "loss": 0.0581, "step": 89250 }, { "epoch": 8.99, "learning_rate": 2.7613991728033894e-05, "loss": 0.0561, "step": 89300 }, { "epoch": 8.99, "learning_rate": 2.7601382023605365e-05, "loss": 0.0599, "step": 89350 }, { "epoch": 9.0, "learning_rate": 2.7588772319176844e-05, "loss": 0.0718, "step": 89400 }, { "epoch": 9.0, "learning_rate": 2.7576162614748312e-05, "loss": 0.056, "step": 89450 }, { "epoch": 9.01, "learning_rate": 2.7563552910319783e-05, "loss": 0.0366, "step": 89500 }, { "epoch": 9.01, "eval_cer": 0.06553589997447058, "eval_loss": 0.4871068000793457, "eval_runtime": 1438.38, "eval_samples_per_second": 5.67, "eval_steps_per_second": 0.709, "step": 89500 }, { "epoch": 9.01, "learning_rate": 2.7550943205891255e-05, "loss": 0.037, "step": 89550 }, { "epoch": 9.02, "learning_rate": 2.7538333501462726e-05, "loss": 0.035, "step": 89600 }, { "epoch": 9.02, "learning_rate": 2.7525723797034198e-05, "loss": 0.0447, "step": 89650 }, { "epoch": 9.03, "learning_rate": 2.751311409260567e-05, "loss": 0.0404, "step": 89700 }, { "epoch": 9.03, "learning_rate": 2.7500504388177144e-05, "loss": 0.037, "step": 89750 }, { "epoch": 9.04, "learning_rate": 2.7487894683748616e-05, "loss": 0.0388, "step": 89800 }, { "epoch": 9.04, "learning_rate": 2.7475284979320087e-05, "loss": 0.0375, "step": 89850 }, { "epoch": 9.05, "learning_rate": 2.746267527489156e-05, "loss": 0.0373, "step": 89900 }, { "epoch": 9.05, "learning_rate": 2.745006557046303e-05, "loss": 0.037, "step": 89950 }, { "epoch": 9.06, "learning_rate": 2.74374558660345e-05, "loss": 0.0369, "step": 90000 }, { "epoch": 9.06, "eval_cer": 0.06449915628366304, "eval_loss": 0.4966026246547699, "eval_runtime": 1353.5139, "eval_samples_per_second": 6.025, "eval_steps_per_second": 0.754, "step": 90000 }, { "epoch": 9.06, "learning_rate": 2.7425098355694546e-05, "loss": 0.0397, "step": 90050 }, { "epoch": 9.07, "learning_rate": 2.7412488651266018e-05, "loss": 0.043, "step": 90100 }, { "epoch": 9.07, "learning_rate": 2.7399878946837486e-05, "loss": 0.0393, "step": 90150 }, { "epoch": 9.08, "learning_rate": 2.7387269242408957e-05, "loss": 0.0376, "step": 90200 }, { "epoch": 9.08, "learning_rate": 2.737465953798043e-05, "loss": 0.0364, "step": 90250 }, { "epoch": 9.09, "learning_rate": 2.73620498335519e-05, "loss": 0.0403, "step": 90300 }, { "epoch": 9.09, "learning_rate": 2.734944012912338e-05, "loss": 0.0407, "step": 90350 }, { "epoch": 9.1, "learning_rate": 2.7336830424694847e-05, "loss": 0.0379, "step": 90400 }, { "epoch": 9.1, "learning_rate": 2.7324220720266318e-05, "loss": 0.038, "step": 90450 }, { "epoch": 9.11, "learning_rate": 2.731161101583779e-05, "loss": 0.0413, "step": 90500 }, { "epoch": 9.11, "eval_cer": 0.06528371907670658, "eval_loss": 0.49570631980895996, "eval_runtime": 1354.3669, "eval_samples_per_second": 6.021, "eval_steps_per_second": 0.753, "step": 90500 }, { "epoch": 9.11, "learning_rate": 2.729900131140926e-05, "loss": 0.0384, "step": 90550 }, { "epoch": 9.12, "learning_rate": 2.7286391606980732e-05, "loss": 0.0352, "step": 90600 }, { "epoch": 9.12, "learning_rate": 2.7273781902552204e-05, "loss": 0.0417, "step": 90650 }, { "epoch": 9.13, "learning_rate": 2.726117219812368e-05, "loss": 0.0374, "step": 90700 }, { "epoch": 9.13, "learning_rate": 2.724856249369515e-05, "loss": 0.0384, "step": 90750 }, { "epoch": 9.14, "learning_rate": 2.7235952789266622e-05, "loss": 0.0398, "step": 90800 }, { "epoch": 9.14, "learning_rate": 2.7223343084838093e-05, "loss": 0.0387, "step": 90850 }, { "epoch": 9.15, "learning_rate": 2.7210733380409565e-05, "loss": 0.0386, "step": 90900 }, { "epoch": 9.15, "learning_rate": 2.7198123675981036e-05, "loss": 0.0331, "step": 90950 }, { "epoch": 9.16, "learning_rate": 2.7185513971552508e-05, "loss": 0.0372, "step": 91000 }, { "epoch": 9.16, "eval_cer": 0.06529617245437394, "eval_loss": 0.49974343180656433, "eval_runtime": 1349.0511, "eval_samples_per_second": 6.045, "eval_steps_per_second": 0.756, "step": 91000 }, { "epoch": 9.16, "learning_rate": 2.717290426712398e-05, "loss": 0.0432, "step": 91050 }, { "epoch": 9.17, "learning_rate": 2.7160294562695454e-05, "loss": 0.0352, "step": 91100 }, { "epoch": 9.17, "learning_rate": 2.7147684858266926e-05, "loss": 0.0385, "step": 91150 }, { "epoch": 9.18, "learning_rate": 2.7135075153838397e-05, "loss": 0.0398, "step": 91200 }, { "epoch": 9.18, "learning_rate": 2.712246544940987e-05, "loss": 0.0502, "step": 91250 }, { "epoch": 9.19, "learning_rate": 2.710985574498134e-05, "loss": 0.0394, "step": 91300 }, { "epoch": 9.19, "learning_rate": 2.7097246040552808e-05, "loss": 0.0419, "step": 91350 }, { "epoch": 9.2, "learning_rate": 2.708463633612428e-05, "loss": 0.0387, "step": 91400 }, { "epoch": 9.2, "learning_rate": 2.707202663169575e-05, "loss": 0.0498, "step": 91450 }, { "epoch": 9.21, "learning_rate": 2.705941692726723e-05, "loss": 0.0453, "step": 91500 }, { "epoch": 9.21, "eval_cer": 0.06704275867222087, "eval_loss": 0.4991973638534546, "eval_runtime": 1333.7477, "eval_samples_per_second": 6.114, "eval_steps_per_second": 0.765, "step": 91500 }, { "epoch": 9.21, "learning_rate": 2.70468072228387e-05, "loss": 0.0437, "step": 91550 }, { "epoch": 9.22, "learning_rate": 2.703419751841017e-05, "loss": 0.0329, "step": 91600 }, { "epoch": 9.22, "learning_rate": 2.702158781398164e-05, "loss": 0.0386, "step": 91650 }, { "epoch": 9.23, "learning_rate": 2.7008978109553112e-05, "loss": 0.0396, "step": 91700 }, { "epoch": 9.23, "learning_rate": 2.6996368405124583e-05, "loss": 0.0435, "step": 91750 }, { "epoch": 9.24, "learning_rate": 2.6983758700696055e-05, "loss": 0.0448, "step": 91800 }, { "epoch": 9.24, "learning_rate": 2.697114899626753e-05, "loss": 0.0383, "step": 91850 }, { "epoch": 9.25, "learning_rate": 2.6958539291839e-05, "loss": 0.0458, "step": 91900 }, { "epoch": 9.25, "learning_rate": 2.6945929587410473e-05, "loss": 0.0391, "step": 91950 }, { "epoch": 9.26, "learning_rate": 2.6933319882981944e-05, "loss": 0.0412, "step": 92000 }, { "epoch": 9.26, "eval_cer": 0.06471397704842496, "eval_loss": 0.49735331535339355, "eval_runtime": 1308.3972, "eval_samples_per_second": 6.233, "eval_steps_per_second": 0.78, "step": 92000 }, { "epoch": 9.26, "learning_rate": 2.6920710178553416e-05, "loss": 0.0457, "step": 92050 }, { "epoch": 9.27, "learning_rate": 2.6908100474124887e-05, "loss": 0.0364, "step": 92100 }, { "epoch": 9.27, "learning_rate": 2.689549076969636e-05, "loss": 0.0415, "step": 92150 }, { "epoch": 9.28, "learning_rate": 2.688288106526783e-05, "loss": 0.0388, "step": 92200 }, { "epoch": 9.28, "learning_rate": 2.6870271360839305e-05, "loss": 0.0393, "step": 92250 }, { "epoch": 9.29, "learning_rate": 2.6857661656410776e-05, "loss": 0.0409, "step": 92300 }, { "epoch": 9.29, "learning_rate": 2.6845051951982248e-05, "loss": 0.0378, "step": 92350 }, { "epoch": 9.3, "learning_rate": 2.683244224755372e-05, "loss": 0.0407, "step": 92400 }, { "epoch": 9.3, "learning_rate": 2.681983254312519e-05, "loss": 0.0479, "step": 92450 }, { "epoch": 9.31, "learning_rate": 2.680722283869666e-05, "loss": 0.0391, "step": 92500 }, { "epoch": 9.31, "eval_cer": 0.06420650190848012, "eval_loss": 0.49970176815986633, "eval_runtime": 1317.7698, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 92500 }, { "epoch": 9.31, "learning_rate": 2.679461313426813e-05, "loss": 0.0479, "step": 92550 }, { "epoch": 9.32, "learning_rate": 2.678200342983961e-05, "loss": 0.0447, "step": 92600 }, { "epoch": 9.32, "learning_rate": 2.676939372541108e-05, "loss": 0.0517, "step": 92650 }, { "epoch": 9.33, "learning_rate": 2.6756784020982552e-05, "loss": 0.0409, "step": 92700 }, { "epoch": 9.33, "learning_rate": 2.6744174316554023e-05, "loss": 0.0369, "step": 92750 }, { "epoch": 9.34, "learning_rate": 2.673156461212549e-05, "loss": 0.0476, "step": 92800 }, { "epoch": 9.34, "learning_rate": 2.6718954907696963e-05, "loss": 0.044, "step": 92850 }, { "epoch": 9.35, "learning_rate": 2.6706345203268434e-05, "loss": 0.0434, "step": 92900 }, { "epoch": 9.35, "learning_rate": 2.6693735498839906e-05, "loss": 0.0423, "step": 92950 }, { "epoch": 9.36, "learning_rate": 2.6681125794411384e-05, "loss": 0.0445, "step": 93000 }, { "epoch": 9.36, "eval_cer": 0.06477313059234491, "eval_loss": 0.4990372657775879, "eval_runtime": 1317.9578, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 93000 }, { "epoch": 9.36, "learning_rate": 2.6668516089982852e-05, "loss": 0.0395, "step": 93050 }, { "epoch": 9.37, "learning_rate": 2.6655906385554324e-05, "loss": 0.0481, "step": 93100 }, { "epoch": 9.37, "learning_rate": 2.6643296681125795e-05, "loss": 0.0447, "step": 93150 }, { "epoch": 9.38, "learning_rate": 2.6630686976697267e-05, "loss": 0.0445, "step": 93200 }, { "epoch": 9.38, "learning_rate": 2.6618077272268738e-05, "loss": 0.0374, "step": 93250 }, { "epoch": 9.39, "learning_rate": 2.660546756784021e-05, "loss": 0.0468, "step": 93300 }, { "epoch": 9.39, "learning_rate": 2.6592857863411684e-05, "loss": 0.0557, "step": 93350 }, { "epoch": 9.4, "learning_rate": 2.6580248158983156e-05, "loss": 0.0432, "step": 93400 }, { "epoch": 9.4, "learning_rate": 2.6567638454554627e-05, "loss": 0.0413, "step": 93450 }, { "epoch": 9.41, "learning_rate": 2.65550287501261e-05, "loss": 0.0427, "step": 93500 }, { "epoch": 9.41, "eval_cer": 0.06407885478738971, "eval_loss": 0.5009150505065918, "eval_runtime": 1326.7151, "eval_samples_per_second": 6.147, "eval_steps_per_second": 0.769, "step": 93500 }, { "epoch": 9.41, "learning_rate": 2.654241904569757e-05, "loss": 0.0448, "step": 93550 }, { "epoch": 9.42, "learning_rate": 2.6529809341269042e-05, "loss": 0.041, "step": 93600 }, { "epoch": 9.42, "learning_rate": 2.6517199636840513e-05, "loss": 0.0427, "step": 93650 }, { "epoch": 9.43, "learning_rate": 2.6504842126500558e-05, "loss": 0.0486, "step": 93700 }, { "epoch": 9.43, "learning_rate": 2.6492232422072026e-05, "loss": 0.0483, "step": 93750 }, { "epoch": 9.44, "learning_rate": 2.6479622717643498e-05, "loss": 0.0394, "step": 93800 }, { "epoch": 9.44, "learning_rate": 2.646701301321497e-05, "loss": 0.0423, "step": 93850 }, { "epoch": 9.45, "learning_rate": 2.645440330878644e-05, "loss": 0.0412, "step": 93900 }, { "epoch": 9.45, "learning_rate": 2.644179360435792e-05, "loss": 0.0464, "step": 93950 }, { "epoch": 9.46, "learning_rate": 2.6429183899929387e-05, "loss": 0.0448, "step": 94000 }, { "epoch": 9.46, "eval_cer": 0.06471397704842496, "eval_loss": 0.49925780296325684, "eval_runtime": 1342.523, "eval_samples_per_second": 6.074, "eval_steps_per_second": 0.76, "step": 94000 }, { "epoch": 9.46, "learning_rate": 2.641657419550086e-05, "loss": 0.0382, "step": 94050 }, { "epoch": 9.47, "learning_rate": 2.640396449107233e-05, "loss": 0.0447, "step": 94100 }, { "epoch": 9.47, "learning_rate": 2.63913547866438e-05, "loss": 0.0434, "step": 94150 }, { "epoch": 9.48, "learning_rate": 2.6378745082215273e-05, "loss": 0.0449, "step": 94200 }, { "epoch": 9.48, "learning_rate": 2.6366135377786744e-05, "loss": 0.0442, "step": 94250 }, { "epoch": 9.49, "learning_rate": 2.635352567335822e-05, "loss": 0.0458, "step": 94300 }, { "epoch": 9.49, "learning_rate": 2.634091596892969e-05, "loss": 0.0423, "step": 94350 }, { "epoch": 9.5, "learning_rate": 2.6328306264501162e-05, "loss": 0.0504, "step": 94400 }, { "epoch": 9.5, "learning_rate": 2.6315696560072634e-05, "loss": 0.0418, "step": 94450 }, { "epoch": 9.51, "learning_rate": 2.6303086855644105e-05, "loss": 0.0411, "step": 94500 }, { "epoch": 9.51, "eval_cer": 0.06508135168961202, "eval_loss": 0.502021849155426, "eval_runtime": 1335.0893, "eval_samples_per_second": 6.108, "eval_steps_per_second": 0.764, "step": 94500 }, { "epoch": 9.51, "learning_rate": 2.6290477151215577e-05, "loss": 0.0433, "step": 94550 }, { "epoch": 9.52, "learning_rate": 2.6277867446787048e-05, "loss": 0.0466, "step": 94600 }, { "epoch": 9.52, "learning_rate": 2.6265257742358516e-05, "loss": 0.0459, "step": 94650 }, { "epoch": 9.53, "learning_rate": 2.6252648037929994e-05, "loss": 0.0449, "step": 94700 }, { "epoch": 9.53, "learning_rate": 2.6240038333501466e-05, "loss": 0.0425, "step": 94750 }, { "epoch": 9.54, "learning_rate": 2.6227428629072937e-05, "loss": 0.044, "step": 94800 }, { "epoch": 9.54, "learning_rate": 2.621481892464441e-05, "loss": 0.0443, "step": 94850 }, { "epoch": 9.55, "learning_rate": 2.6202209220215877e-05, "loss": 0.0464, "step": 94900 }, { "epoch": 9.55, "learning_rate": 2.618959951578735e-05, "loss": 0.044, "step": 94950 }, { "epoch": 9.56, "learning_rate": 2.617698981135882e-05, "loss": 0.0432, "step": 95000 }, { "epoch": 9.56, "eval_cer": 0.06377686037895629, "eval_loss": 0.5018033385276794, "eval_runtime": 1378.4097, "eval_samples_per_second": 5.916, "eval_steps_per_second": 0.74, "step": 95000 }, { "epoch": 9.56, "learning_rate": 2.6164380106930298e-05, "loss": 0.044, "step": 95050 }, { "epoch": 9.57, "learning_rate": 2.615177040250177e-05, "loss": 0.0423, "step": 95100 }, { "epoch": 9.57, "learning_rate": 2.613916069807324e-05, "loss": 0.0398, "step": 95150 }, { "epoch": 9.58, "learning_rate": 2.612655099364471e-05, "loss": 0.0402, "step": 95200 }, { "epoch": 9.58, "learning_rate": 2.611394128921618e-05, "loss": 0.0444, "step": 95250 }, { "epoch": 9.59, "learning_rate": 2.6101331584787652e-05, "loss": 0.0403, "step": 95300 }, { "epoch": 9.59, "learning_rate": 2.6088721880359124e-05, "loss": 0.0439, "step": 95350 }, { "epoch": 9.6, "learning_rate": 2.6076112175930595e-05, "loss": 0.0456, "step": 95400 }, { "epoch": 9.6, "learning_rate": 2.606350247150207e-05, "loss": 0.0444, "step": 95450 }, { "epoch": 9.61, "learning_rate": 2.605089276707354e-05, "loss": 0.0445, "step": 95500 }, { "epoch": 9.61, "eval_cer": 0.06462680340475345, "eval_loss": 0.5058117508888245, "eval_runtime": 1465.9919, "eval_samples_per_second": 5.563, "eval_steps_per_second": 0.696, "step": 95500 }, { "epoch": 9.61, "learning_rate": 2.6038283062645013e-05, "loss": 0.0429, "step": 95550 }, { "epoch": 9.62, "learning_rate": 2.6025673358216485e-05, "loss": 0.0425, "step": 95600 }, { "epoch": 9.62, "learning_rate": 2.6013063653787956e-05, "loss": 0.0441, "step": 95650 }, { "epoch": 9.63, "learning_rate": 2.6000453949359427e-05, "loss": 0.0425, "step": 95700 }, { "epoch": 9.63, "learning_rate": 2.59878442449309e-05, "loss": 0.0411, "step": 95750 }, { "epoch": 9.64, "learning_rate": 2.5975234540502374e-05, "loss": 0.0451, "step": 95800 }, { "epoch": 9.64, "learning_rate": 2.5962624836073845e-05, "loss": 0.0447, "step": 95850 }, { "epoch": 9.65, "learning_rate": 2.5950015131645317e-05, "loss": 0.0446, "step": 95900 }, { "epoch": 9.65, "learning_rate": 2.5937405427216788e-05, "loss": 0.0457, "step": 95950 }, { "epoch": 9.66, "learning_rate": 2.592479572278826e-05, "loss": 0.0441, "step": 96000 }, { "epoch": 9.66, "eval_cer": 0.06375195362362157, "eval_loss": 0.5010879635810852, "eval_runtime": 1654.3618, "eval_samples_per_second": 4.929, "eval_steps_per_second": 0.617, "step": 96000 }, { "epoch": 9.66, "learning_rate": 2.591218601835973e-05, "loss": 0.0465, "step": 96050 }, { "epoch": 9.67, "learning_rate": 2.58995763139312e-05, "loss": 0.0431, "step": 96100 }, { "epoch": 9.67, "learning_rate": 2.588696660950267e-05, "loss": 0.0467, "step": 96150 }, { "epoch": 9.68, "learning_rate": 2.587435690507415e-05, "loss": 0.0449, "step": 96200 }, { "epoch": 9.69, "learning_rate": 2.586174720064562e-05, "loss": 0.0477, "step": 96250 }, { "epoch": 9.69, "learning_rate": 2.5849137496217092e-05, "loss": 0.0462, "step": 96300 }, { "epoch": 9.7, "learning_rate": 2.583652779178856e-05, "loss": 0.0435, "step": 96350 }, { "epoch": 9.7, "learning_rate": 2.582391808736003e-05, "loss": 0.044, "step": 96400 }, { "epoch": 9.71, "learning_rate": 2.5811308382931503e-05, "loss": 0.0384, "step": 96450 }, { "epoch": 9.71, "learning_rate": 2.5798698678502975e-05, "loss": 0.0443, "step": 96500 }, { "epoch": 9.71, "eval_cer": 0.06466105019333869, "eval_loss": 0.5021639466285706, "eval_runtime": 1668.3985, "eval_samples_per_second": 4.888, "eval_steps_per_second": 0.611, "step": 96500 }, { "epoch": 9.72, "learning_rate": 2.5786088974074453e-05, "loss": 0.0527, "step": 96550 }, { "epoch": 9.72, "learning_rate": 2.577347926964592e-05, "loss": 0.045, "step": 96600 }, { "epoch": 9.73, "learning_rate": 2.5760869565217392e-05, "loss": 0.0445, "step": 96650 }, { "epoch": 9.73, "learning_rate": 2.5748259860788864e-05, "loss": 0.0454, "step": 96700 }, { "epoch": 9.74, "learning_rate": 2.573590235044891e-05, "loss": 0.0488, "step": 96750 }, { "epoch": 9.74, "learning_rate": 2.572329264602038e-05, "loss": 0.04, "step": 96800 }, { "epoch": 9.75, "learning_rate": 2.571068294159185e-05, "loss": 0.0442, "step": 96850 }, { "epoch": 9.75, "learning_rate": 2.5698073237163323e-05, "loss": 0.0491, "step": 96900 }, { "epoch": 9.76, "learning_rate": 2.5685463532734795e-05, "loss": 0.0448, "step": 96950 }, { "epoch": 9.76, "learning_rate": 2.5672853828306266e-05, "loss": 0.0408, "step": 97000 }, { "epoch": 9.76, "eval_cer": 0.06480115069209647, "eval_loss": 0.5022784471511841, "eval_runtime": 1651.2684, "eval_samples_per_second": 4.939, "eval_steps_per_second": 0.618, "step": 97000 }, { "epoch": 9.77, "learning_rate": 2.5660244123877734e-05, "loss": 0.0441, "step": 97050 }, { "epoch": 9.77, "learning_rate": 2.5647634419449206e-05, "loss": 0.0452, "step": 97100 }, { "epoch": 9.78, "learning_rate": 2.5635024715020684e-05, "loss": 0.0418, "step": 97150 }, { "epoch": 9.78, "learning_rate": 2.5622667204680722e-05, "loss": 0.0507, "step": 97200 }, { "epoch": 9.79, "learning_rate": 2.5610057500252193e-05, "loss": 0.0456, "step": 97250 }, { "epoch": 9.79, "learning_rate": 2.5597447795823665e-05, "loss": 0.0517, "step": 97300 }, { "epoch": 9.8, "learning_rate": 2.558483809139514e-05, "loss": 0.0437, "step": 97350 }, { "epoch": 9.8, "learning_rate": 2.557222838696661e-05, "loss": 0.0402, "step": 97400 }, { "epoch": 9.81, "learning_rate": 2.5559618682538083e-05, "loss": 0.0513, "step": 97450 }, { "epoch": 9.81, "learning_rate": 2.5547008978109554e-05, "loss": 0.0409, "step": 97500 }, { "epoch": 9.81, "eval_cer": 0.06500974476802471, "eval_loss": 0.5006101727485657, "eval_runtime": 1634.6414, "eval_samples_per_second": 4.989, "eval_steps_per_second": 0.624, "step": 97500 }, { "epoch": 9.82, "learning_rate": 2.5534399273681026e-05, "loss": 0.0411, "step": 97550 }, { "epoch": 9.82, "learning_rate": 2.5521789569252497e-05, "loss": 0.0499, "step": 97600 }, { "epoch": 9.83, "learning_rate": 2.550917986482397e-05, "loss": 0.0466, "step": 97650 }, { "epoch": 9.83, "learning_rate": 2.5496570160395443e-05, "loss": 0.0496, "step": 97700 }, { "epoch": 9.84, "learning_rate": 2.5483960455966915e-05, "loss": 0.0482, "step": 97750 }, { "epoch": 9.84, "learning_rate": 2.5471350751538386e-05, "loss": 0.0498, "step": 97800 }, { "epoch": 9.85, "learning_rate": 2.5458741047109858e-05, "loss": 0.0504, "step": 97850 }, { "epoch": 9.85, "learning_rate": 2.544613134268133e-05, "loss": 0.0466, "step": 97900 }, { "epoch": 9.86, "learning_rate": 2.54335216382528e-05, "loss": 0.0462, "step": 97950 }, { "epoch": 9.86, "learning_rate": 2.5420911933824272e-05, "loss": 0.0465, "step": 98000 }, { "epoch": 9.86, "eval_cer": 0.06428744886331796, "eval_loss": 0.5037218332290649, "eval_runtime": 1402.0897, "eval_samples_per_second": 5.816, "eval_steps_per_second": 0.727, "step": 98000 }, { "epoch": 9.87, "learning_rate": 2.540830222939574e-05, "loss": 0.0427, "step": 98050 }, { "epoch": 9.87, "learning_rate": 2.539569252496722e-05, "loss": 0.0503, "step": 98100 }, { "epoch": 9.88, "learning_rate": 2.538308282053869e-05, "loss": 0.041, "step": 98150 }, { "epoch": 9.88, "learning_rate": 2.537047311611016e-05, "loss": 0.0495, "step": 98200 }, { "epoch": 9.89, "learning_rate": 2.5357863411681633e-05, "loss": 0.0485, "step": 98250 }, { "epoch": 9.89, "learning_rate": 2.53452537072531e-05, "loss": 0.0499, "step": 98300 }, { "epoch": 9.9, "learning_rate": 2.5332644002824573e-05, "loss": 0.0443, "step": 98350 }, { "epoch": 9.9, "learning_rate": 2.5320034298396044e-05, "loss": 0.0469, "step": 98400 }, { "epoch": 9.91, "learning_rate": 2.5307424593967516e-05, "loss": 0.0415, "step": 98450 }, { "epoch": 9.91, "learning_rate": 2.5294814889538994e-05, "loss": 0.048, "step": 98500 }, { "epoch": 9.91, "eval_cer": 0.06397922776605085, "eval_loss": 0.4986913204193115, "eval_runtime": 1630.2639, "eval_samples_per_second": 5.002, "eval_steps_per_second": 0.626, "step": 98500 }, { "epoch": 9.92, "learning_rate": 2.5282205185110462e-05, "loss": 0.0506, "step": 98550 }, { "epoch": 9.92, "learning_rate": 2.5269595480681933e-05, "loss": 0.048, "step": 98600 }, { "epoch": 9.93, "learning_rate": 2.5256985776253405e-05, "loss": 0.042, "step": 98650 }, { "epoch": 9.93, "learning_rate": 2.5244376071824876e-05, "loss": 0.0442, "step": 98700 }, { "epoch": 9.94, "learning_rate": 2.5231766367396348e-05, "loss": 0.0465, "step": 98750 }, { "epoch": 9.94, "learning_rate": 2.521915666296782e-05, "loss": 0.0407, "step": 98800 }, { "epoch": 9.95, "learning_rate": 2.5206546958539294e-05, "loss": 0.0516, "step": 98850 }, { "epoch": 9.95, "learning_rate": 2.5193937254110766e-05, "loss": 0.0454, "step": 98900 }, { "epoch": 9.96, "learning_rate": 2.5181327549682237e-05, "loss": 0.0457, "step": 98950 }, { "epoch": 9.96, "learning_rate": 2.516871784525371e-05, "loss": 0.0533, "step": 99000 }, { "epoch": 9.96, "eval_cer": 0.06412244160922546, "eval_loss": 0.5008233189582825, "eval_runtime": 1561.8059, "eval_samples_per_second": 5.222, "eval_steps_per_second": 0.653, "step": 99000 }, { "epoch": 9.97, "learning_rate": 2.515610814082518e-05, "loss": 0.044, "step": 99050 }, { "epoch": 9.97, "learning_rate": 2.5143498436396652e-05, "loss": 0.0549, "step": 99100 }, { "epoch": 9.98, "learning_rate": 2.5130888731968123e-05, "loss": 0.0403, "step": 99150 }, { "epoch": 9.98, "learning_rate": 2.511827902753959e-05, "loss": 0.0462, "step": 99200 }, { "epoch": 9.99, "learning_rate": 2.510566932311107e-05, "loss": 0.0388, "step": 99250 }, { "epoch": 9.99, "learning_rate": 2.509305961868254e-05, "loss": 0.0489, "step": 99300 }, { "epoch": 10.0, "learning_rate": 2.5080449914254013e-05, "loss": 0.0502, "step": 99350 }, { "epoch": 10.0, "learning_rate": 2.5067840209825484e-05, "loss": 0.0354, "step": 99400 }, { "epoch": 10.01, "learning_rate": 2.5055230505396955e-05, "loss": 0.0257, "step": 99450 }, { "epoch": 10.01, "learning_rate": 2.5042620800968424e-05, "loss": 0.0267, "step": 99500 }, { "epoch": 10.01, "eval_cer": 0.06379865378987416, "eval_loss": 0.5068789124488831, "eval_runtime": 1615.0384, "eval_samples_per_second": 5.049, "eval_steps_per_second": 0.632, "step": 99500 }, { "epoch": 10.02, "learning_rate": 2.5030011096539895e-05, "loss": 0.0255, "step": 99550 }, { "epoch": 10.02, "learning_rate": 2.5017401392111373e-05, "loss": 0.0269, "step": 99600 }, { "epoch": 10.03, "learning_rate": 2.5004791687682845e-05, "loss": 0.0232, "step": 99650 }, { "epoch": 10.03, "learning_rate": 2.4992181983254316e-05, "loss": 0.027, "step": 99700 }, { "epoch": 10.04, "learning_rate": 2.4979572278825784e-05, "loss": 0.0255, "step": 99750 }, { "epoch": 10.04, "learning_rate": 2.4966962574397256e-05, "loss": 0.0273, "step": 99800 }, { "epoch": 10.05, "learning_rate": 2.4954352869968727e-05, "loss": 0.0281, "step": 99850 }, { "epoch": 10.05, "learning_rate": 2.4941743165540202e-05, "loss": 0.0309, "step": 99900 }, { "epoch": 10.06, "learning_rate": 2.4929133461111674e-05, "loss": 0.0249, "step": 99950 }, { "epoch": 10.06, "learning_rate": 2.4916523756683145e-05, "loss": 0.031, "step": 100000 }, { "epoch": 10.06, "eval_cer": 0.06322268507275886, "eval_loss": 0.510090172290802, "eval_runtime": 1306.6194, "eval_samples_per_second": 6.241, "eval_steps_per_second": 0.781, "step": 100000 }, { "epoch": 10.07, "learning_rate": 2.4903914052254617e-05, "loss": 0.0262, "step": 100050 }, { "epoch": 10.07, "learning_rate": 2.4891304347826088e-05, "loss": 0.0265, "step": 100100 }, { "epoch": 10.08, "learning_rate": 2.487869464339756e-05, "loss": 0.0303, "step": 100150 }, { "epoch": 10.08, "learning_rate": 2.486608493896903e-05, "loss": 0.0298, "step": 100200 }, { "epoch": 10.09, "learning_rate": 2.4853475234540506e-05, "loss": 0.0288, "step": 100250 }, { "epoch": 10.09, "learning_rate": 2.4840865530111977e-05, "loss": 0.0307, "step": 100300 }, { "epoch": 10.1, "learning_rate": 2.4828255825683446e-05, "loss": 0.0327, "step": 100350 }, { "epoch": 10.1, "learning_rate": 2.4815646121254917e-05, "loss": 0.0295, "step": 100400 }, { "epoch": 10.11, "learning_rate": 2.4803036416826392e-05, "loss": 0.0298, "step": 100450 }, { "epoch": 10.11, "learning_rate": 2.4790426712397863e-05, "loss": 0.0388, "step": 100500 }, { "epoch": 10.11, "eval_cer": 0.06427499548565059, "eval_loss": 0.5115137100219727, "eval_runtime": 1317.9587, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 100500 }, { "epoch": 10.12, "learning_rate": 2.4777817007969335e-05, "loss": 0.0285, "step": 100550 }, { "epoch": 10.12, "learning_rate": 2.4765207303540806e-05, "loss": 0.03, "step": 100600 }, { "epoch": 10.13, "learning_rate": 2.4752597599112278e-05, "loss": 0.0329, "step": 100650 }, { "epoch": 10.13, "learning_rate": 2.473998789468375e-05, "loss": 0.0304, "step": 100700 }, { "epoch": 10.14, "learning_rate": 2.472737819025522e-05, "loss": 0.0261, "step": 100750 }, { "epoch": 10.14, "learning_rate": 2.4714768485826692e-05, "loss": 0.0275, "step": 100800 }, { "epoch": 10.15, "learning_rate": 2.4702158781398167e-05, "loss": 0.029, "step": 100850 }, { "epoch": 10.15, "learning_rate": 2.4689549076969635e-05, "loss": 0.0276, "step": 100900 }, { "epoch": 10.16, "learning_rate": 2.4676939372541107e-05, "loss": 0.029, "step": 100950 }, { "epoch": 10.16, "learning_rate": 2.466432966811258e-05, "loss": 0.0287, "step": 101000 }, { "epoch": 10.16, "eval_cer": 0.06405706137647184, "eval_loss": 0.5145584940910339, "eval_runtime": 1312.5526, "eval_samples_per_second": 6.213, "eval_steps_per_second": 0.777, "step": 101000 }, { "epoch": 10.17, "learning_rate": 2.4651719963684053e-05, "loss": 0.0328, "step": 101050 }, { "epoch": 10.17, "learning_rate": 2.4639110259255525e-05, "loss": 0.0326, "step": 101100 }, { "epoch": 10.18, "learning_rate": 2.4626500554826996e-05, "loss": 0.0284, "step": 101150 }, { "epoch": 10.18, "learning_rate": 2.4613890850398468e-05, "loss": 0.0313, "step": 101200 }, { "epoch": 10.19, "learning_rate": 2.460128114596994e-05, "loss": 0.0304, "step": 101250 }, { "epoch": 10.19, "learning_rate": 2.458867144154141e-05, "loss": 0.0318, "step": 101300 }, { "epoch": 10.2, "learning_rate": 2.4576061737112882e-05, "loss": 0.0356, "step": 101350 }, { "epoch": 10.2, "learning_rate": 2.4563452032684357e-05, "loss": 0.0308, "step": 101400 }, { "epoch": 10.21, "learning_rate": 2.455084232825583e-05, "loss": 0.0355, "step": 101450 }, { "epoch": 10.21, "learning_rate": 2.4538232623827296e-05, "loss": 0.0294, "step": 101500 }, { "epoch": 10.21, "eval_cer": 0.06385469398937726, "eval_loss": 0.5088115930557251, "eval_runtime": 1317.3091, "eval_samples_per_second": 6.191, "eval_steps_per_second": 0.774, "step": 101500 }, { "epoch": 10.22, "learning_rate": 2.4525622919398768e-05, "loss": 0.0267, "step": 101550 }, { "epoch": 10.22, "learning_rate": 2.4513013214970243e-05, "loss": 0.0325, "step": 101600 }, { "epoch": 10.23, "learning_rate": 2.4500403510541714e-05, "loss": 0.0262, "step": 101650 }, { "epoch": 10.23, "learning_rate": 2.4487793806113186e-05, "loss": 0.0294, "step": 101700 }, { "epoch": 10.24, "learning_rate": 2.4475184101684657e-05, "loss": 0.0299, "step": 101750 }, { "epoch": 10.24, "learning_rate": 2.446257439725613e-05, "loss": 0.0289, "step": 101800 }, { "epoch": 10.25, "learning_rate": 2.44499646928276e-05, "loss": 0.0303, "step": 101850 }, { "epoch": 10.25, "learning_rate": 2.4437354988399072e-05, "loss": 0.034, "step": 101900 }, { "epoch": 10.26, "learning_rate": 2.4424745283970547e-05, "loss": 0.0307, "step": 101950 }, { "epoch": 10.26, "learning_rate": 2.4412135579542018e-05, "loss": 0.0276, "step": 102000 }, { "epoch": 10.26, "eval_cer": 0.06451783635016407, "eval_loss": 0.5178592205047607, "eval_runtime": 1305.3246, "eval_samples_per_second": 6.247, "eval_steps_per_second": 0.781, "step": 102000 }, { "epoch": 10.27, "learning_rate": 2.439952587511349e-05, "loss": 0.0301, "step": 102050 }, { "epoch": 10.27, "learning_rate": 2.4386916170684958e-05, "loss": 0.0361, "step": 102100 }, { "epoch": 10.28, "learning_rate": 2.4374306466256433e-05, "loss": 0.0344, "step": 102150 }, { "epoch": 10.28, "learning_rate": 2.4361696761827904e-05, "loss": 0.0322, "step": 102200 }, { "epoch": 10.29, "learning_rate": 2.4349087057399376e-05, "loss": 0.0295, "step": 102250 }, { "epoch": 10.29, "learning_rate": 2.4336477352970847e-05, "loss": 0.0338, "step": 102300 }, { "epoch": 10.3, "learning_rate": 2.432386764854232e-05, "loss": 0.03, "step": 102350 }, { "epoch": 10.3, "learning_rate": 2.431125794411379e-05, "loss": 0.0308, "step": 102400 }, { "epoch": 10.31, "learning_rate": 2.429864823968526e-05, "loss": 0.0348, "step": 102450 }, { "epoch": 10.31, "learning_rate": 2.4286038535256736e-05, "loss": 0.0333, "step": 102500 }, { "epoch": 10.31, "eval_cer": 0.0644493427729936, "eval_loss": 0.5134919881820679, "eval_runtime": 1320.6923, "eval_samples_per_second": 6.175, "eval_steps_per_second": 0.772, "step": 102500 }, { "epoch": 10.32, "learning_rate": 2.4273428830828208e-05, "loss": 0.0335, "step": 102550 }, { "epoch": 10.32, "learning_rate": 2.426081912639968e-05, "loss": 0.0298, "step": 102600 }, { "epoch": 10.33, "learning_rate": 2.424820942197115e-05, "loss": 0.0331, "step": 102650 }, { "epoch": 10.33, "learning_rate": 2.4235599717542622e-05, "loss": 0.0345, "step": 102700 }, { "epoch": 10.34, "learning_rate": 2.4222990013114094e-05, "loss": 0.0298, "step": 102750 }, { "epoch": 10.34, "learning_rate": 2.4210380308685565e-05, "loss": 0.0312, "step": 102800 }, { "epoch": 10.35, "learning_rate": 2.4197770604257037e-05, "loss": 0.0307, "step": 102850 }, { "epoch": 10.35, "learning_rate": 2.418516089982851e-05, "loss": 0.0273, "step": 102900 }, { "epoch": 10.36, "learning_rate": 2.417255119539998e-05, "loss": 0.0328, "step": 102950 }, { "epoch": 10.36, "learning_rate": 2.415994149097145e-05, "loss": 0.0311, "step": 103000 }, { "epoch": 10.36, "eval_cer": 0.06397300107721716, "eval_loss": 0.5122085213661194, "eval_runtime": 1310.7876, "eval_samples_per_second": 6.221, "eval_steps_per_second": 0.778, "step": 103000 }, { "epoch": 10.37, "learning_rate": 2.4147331786542923e-05, "loss": 0.0311, "step": 103050 }, { "epoch": 10.37, "learning_rate": 2.4134722082114398e-05, "loss": 0.0297, "step": 103100 }, { "epoch": 10.38, "learning_rate": 2.412211237768587e-05, "loss": 0.0353, "step": 103150 }, { "epoch": 10.38, "learning_rate": 2.410950267325734e-05, "loss": 0.0285, "step": 103200 }, { "epoch": 10.39, "learning_rate": 2.4096892968828812e-05, "loss": 0.03, "step": 103250 }, { "epoch": 10.39, "learning_rate": 2.4084283264400283e-05, "loss": 0.031, "step": 103300 }, { "epoch": 10.4, "learning_rate": 2.4071673559971755e-05, "loss": 0.0288, "step": 103350 }, { "epoch": 10.4, "learning_rate": 2.4059063855543226e-05, "loss": 0.0315, "step": 103400 }, { "epoch": 10.41, "learning_rate": 2.40464541511147e-05, "loss": 0.0326, "step": 103450 }, { "epoch": 10.41, "learning_rate": 2.4033844446686173e-05, "loss": 0.0311, "step": 103500 }, { "epoch": 10.41, "eval_cer": 0.06470463701517444, "eval_loss": 0.5157328248023987, "eval_runtime": 1425.3085, "eval_samples_per_second": 5.722, "eval_steps_per_second": 0.716, "step": 103500 }, { "epoch": 10.42, "learning_rate": 2.402123474225764e-05, "loss": 0.0355, "step": 103550 }, { "epoch": 10.42, "learning_rate": 2.4008625037829112e-05, "loss": 0.0303, "step": 103600 }, { "epoch": 10.43, "learning_rate": 2.3996015333400587e-05, "loss": 0.0306, "step": 103650 }, { "epoch": 10.43, "learning_rate": 2.398340562897206e-05, "loss": 0.029, "step": 103700 }, { "epoch": 10.44, "learning_rate": 2.397079592454353e-05, "loss": 0.0345, "step": 103750 }, { "epoch": 10.44, "learning_rate": 2.3958186220115e-05, "loss": 0.0337, "step": 103800 }, { "epoch": 10.45, "learning_rate": 2.3945576515686473e-05, "loss": 0.0313, "step": 103850 }, { "epoch": 10.45, "learning_rate": 2.3932966811257945e-05, "loss": 0.0277, "step": 103900 }, { "epoch": 10.46, "learning_rate": 2.3920357106829416e-05, "loss": 0.0309, "step": 103950 }, { "epoch": 10.46, "learning_rate": 2.390774740240089e-05, "loss": 0.0318, "step": 104000 }, { "epoch": 10.46, "eval_cer": 0.06419093518639593, "eval_loss": 0.5133367776870728, "eval_runtime": 1603.2264, "eval_samples_per_second": 5.087, "eval_steps_per_second": 0.636, "step": 104000 }, { "epoch": 10.47, "learning_rate": 2.3895137697972362e-05, "loss": 0.0304, "step": 104050 }, { "epoch": 10.47, "learning_rate": 2.3882527993543834e-05, "loss": 0.0333, "step": 104100 }, { "epoch": 10.48, "learning_rate": 2.3869918289115302e-05, "loss": 0.0306, "step": 104150 }, { "epoch": 10.48, "learning_rate": 2.3857308584686777e-05, "loss": 0.032, "step": 104200 }, { "epoch": 10.49, "learning_rate": 2.384469888025825e-05, "loss": 0.0291, "step": 104250 }, { "epoch": 10.5, "learning_rate": 2.383208917582972e-05, "loss": 0.0249, "step": 104300 }, { "epoch": 10.5, "learning_rate": 2.381947947140119e-05, "loss": 0.0303, "step": 104350 }, { "epoch": 10.51, "learning_rate": 2.3806869766972663e-05, "loss": 0.0339, "step": 104400 }, { "epoch": 10.51, "learning_rate": 2.3794512256632708e-05, "loss": 0.0367, "step": 104450 }, { "epoch": 10.52, "learning_rate": 2.3781902552204176e-05, "loss": 0.0322, "step": 104500 }, { "epoch": 10.52, "eval_cer": 0.06445245611741045, "eval_loss": 0.5143043994903564, "eval_runtime": 1381.381, "eval_samples_per_second": 5.904, "eval_steps_per_second": 0.738, "step": 104500 }, { "epoch": 10.52, "learning_rate": 2.3769292847775647e-05, "loss": 0.0352, "step": 104550 }, { "epoch": 10.53, "learning_rate": 2.3756683143347122e-05, "loss": 0.0333, "step": 104600 }, { "epoch": 10.53, "learning_rate": 2.3744073438918593e-05, "loss": 0.0305, "step": 104650 }, { "epoch": 10.54, "learning_rate": 2.3731463734490065e-05, "loss": 0.0343, "step": 104700 }, { "epoch": 10.54, "learning_rate": 2.3718854030061536e-05, "loss": 0.0342, "step": 104750 }, { "epoch": 10.55, "learning_rate": 2.3706244325633008e-05, "loss": 0.0367, "step": 104800 }, { "epoch": 10.55, "learning_rate": 2.369363462120448e-05, "loss": 0.0371, "step": 104850 }, { "epoch": 10.56, "learning_rate": 2.368102491677595e-05, "loss": 0.0354, "step": 104900 }, { "epoch": 10.56, "learning_rate": 2.3668415212347426e-05, "loss": 0.0308, "step": 104950 }, { "epoch": 10.57, "learning_rate": 2.3655805507918897e-05, "loss": 0.0315, "step": 105000 }, { "epoch": 10.57, "eval_cer": 0.06369591342411846, "eval_loss": 0.5137677788734436, "eval_runtime": 1349.2026, "eval_samples_per_second": 6.044, "eval_steps_per_second": 0.756, "step": 105000 }, { "epoch": 10.57, "learning_rate": 2.364319580349037e-05, "loss": 0.029, "step": 105050 }, { "epoch": 10.58, "learning_rate": 2.3630586099061837e-05, "loss": 0.0324, "step": 105100 }, { "epoch": 10.58, "learning_rate": 2.3617976394633312e-05, "loss": 0.0318, "step": 105150 }, { "epoch": 10.59, "learning_rate": 2.3605366690204783e-05, "loss": 0.0317, "step": 105200 }, { "epoch": 10.59, "learning_rate": 2.3592756985776255e-05, "loss": 0.0313, "step": 105250 }, { "epoch": 10.6, "learning_rate": 2.3580147281347726e-05, "loss": 0.0281, "step": 105300 }, { "epoch": 10.6, "learning_rate": 2.3567537576919198e-05, "loss": 0.0376, "step": 105350 }, { "epoch": 10.61, "learning_rate": 2.355492787249067e-05, "loss": 0.0298, "step": 105400 }, { "epoch": 10.61, "learning_rate": 2.354231816806214e-05, "loss": 0.0296, "step": 105450 }, { "epoch": 10.62, "learning_rate": 2.3529708463633612e-05, "loss": 0.0316, "step": 105500 }, { "epoch": 10.62, "eval_cer": 0.06404460799880447, "eval_loss": 0.5182689428329468, "eval_runtime": 1337.747, "eval_samples_per_second": 6.096, "eval_steps_per_second": 0.762, "step": 105500 }, { "epoch": 10.62, "learning_rate": 2.3517098759205087e-05, "loss": 0.0298, "step": 105550 }, { "epoch": 10.63, "learning_rate": 2.350448905477656e-05, "loss": 0.0322, "step": 105600 }, { "epoch": 10.63, "learning_rate": 2.349187935034803e-05, "loss": 0.0333, "step": 105650 }, { "epoch": 10.64, "learning_rate": 2.3479269645919498e-05, "loss": 0.031, "step": 105700 }, { "epoch": 10.64, "learning_rate": 2.3466659941490973e-05, "loss": 0.0347, "step": 105750 }, { "epoch": 10.65, "learning_rate": 2.3454050237062444e-05, "loss": 0.0303, "step": 105800 }, { "epoch": 10.65, "learning_rate": 2.3441440532633916e-05, "loss": 0.0334, "step": 105850 }, { "epoch": 10.66, "learning_rate": 2.342883082820539e-05, "loss": 0.0331, "step": 105900 }, { "epoch": 10.66, "learning_rate": 2.341622112377686e-05, "loss": 0.0346, "step": 105950 }, { "epoch": 10.67, "learning_rate": 2.340361141934833e-05, "loss": 0.0307, "step": 106000 }, { "epoch": 10.67, "eval_cer": 0.0627837035099845, "eval_loss": 0.5154832005500793, "eval_runtime": 1323.5537, "eval_samples_per_second": 6.161, "eval_steps_per_second": 0.771, "step": 106000 }, { "epoch": 10.67, "learning_rate": 2.3391001714919802e-05, "loss": 0.0328, "step": 106050 }, { "epoch": 10.68, "learning_rate": 2.3378392010491277e-05, "loss": 0.0343, "step": 106100 }, { "epoch": 10.68, "learning_rate": 2.3365782306062748e-05, "loss": 0.0304, "step": 106150 }, { "epoch": 10.69, "learning_rate": 2.335317260163422e-05, "loss": 0.0341, "step": 106200 }, { "epoch": 10.69, "learning_rate": 2.334056289720569e-05, "loss": 0.0306, "step": 106250 }, { "epoch": 10.7, "learning_rate": 2.3327953192777163e-05, "loss": 0.029, "step": 106300 }, { "epoch": 10.7, "learning_rate": 2.3315343488348634e-05, "loss": 0.0303, "step": 106350 }, { "epoch": 10.71, "learning_rate": 2.3302733783920106e-05, "loss": 0.0351, "step": 106400 }, { "epoch": 10.71, "learning_rate": 2.3290124079491577e-05, "loss": 0.0337, "step": 106450 }, { "epoch": 10.72, "learning_rate": 2.3277514375063052e-05, "loss": 0.0301, "step": 106500 }, { "epoch": 10.72, "eval_cer": 0.06333787881618191, "eval_loss": 0.5214671492576599, "eval_runtime": 1346.485, "eval_samples_per_second": 6.057, "eval_steps_per_second": 0.758, "step": 106500 }, { "epoch": 10.72, "learning_rate": 2.326490467063452e-05, "loss": 0.0307, "step": 106550 }, { "epoch": 10.73, "learning_rate": 2.325229496620599e-05, "loss": 0.0326, "step": 106600 }, { "epoch": 10.73, "learning_rate": 2.3239685261777466e-05, "loss": 0.0299, "step": 106650 }, { "epoch": 10.74, "learning_rate": 2.3227075557348938e-05, "loss": 0.036, "step": 106700 }, { "epoch": 10.74, "learning_rate": 2.321446585292041e-05, "loss": 0.0293, "step": 106750 }, { "epoch": 10.75, "learning_rate": 2.320185614849188e-05, "loss": 0.032, "step": 106800 }, { "epoch": 10.75, "learning_rate": 2.3189246444063352e-05, "loss": 0.0305, "step": 106850 }, { "epoch": 10.76, "learning_rate": 2.3176636739634824e-05, "loss": 0.0311, "step": 106900 }, { "epoch": 10.76, "learning_rate": 2.3164027035206295e-05, "loss": 0.0334, "step": 106950 }, { "epoch": 10.77, "learning_rate": 2.3151417330777767e-05, "loss": 0.0336, "step": 107000 }, { "epoch": 10.77, "eval_cer": 0.0635340195144428, "eval_loss": 0.5136501789093018, "eval_runtime": 1334.5128, "eval_samples_per_second": 6.111, "eval_steps_per_second": 0.764, "step": 107000 }, { "epoch": 10.77, "learning_rate": 2.313880762634924e-05, "loss": 0.0282, "step": 107050 }, { "epoch": 10.78, "learning_rate": 2.3126197921920713e-05, "loss": 0.0281, "step": 107100 }, { "epoch": 10.78, "learning_rate": 2.311358821749218e-05, "loss": 0.0321, "step": 107150 }, { "epoch": 10.79, "learning_rate": 2.3101230707152226e-05, "loss": 0.0375, "step": 107200 }, { "epoch": 10.79, "learning_rate": 2.3088621002723697e-05, "loss": 0.0364, "step": 107250 }, { "epoch": 10.8, "learning_rate": 2.307601129829517e-05, "loss": 0.0287, "step": 107300 }, { "epoch": 10.8, "learning_rate": 2.306340159386664e-05, "loss": 0.0326, "step": 107350 }, { "epoch": 10.81, "learning_rate": 2.3050791889438112e-05, "loss": 0.0316, "step": 107400 }, { "epoch": 10.81, "learning_rate": 2.3038182185009587e-05, "loss": 0.0368, "step": 107450 }, { "epoch": 10.82, "learning_rate": 2.3025572480581055e-05, "loss": 0.0339, "step": 107500 }, { "epoch": 10.82, "eval_cer": 0.0641411216757265, "eval_loss": 0.5151168704032898, "eval_runtime": 1342.6719, "eval_samples_per_second": 6.074, "eval_steps_per_second": 0.76, "step": 107500 }, { "epoch": 10.82, "learning_rate": 2.3012962776152526e-05, "loss": 0.0332, "step": 107550 }, { "epoch": 10.83, "learning_rate": 2.3000353071724e-05, "loss": 0.0373, "step": 107600 }, { "epoch": 10.83, "learning_rate": 2.2987743367295473e-05, "loss": 0.0381, "step": 107650 }, { "epoch": 10.84, "learning_rate": 2.2975133662866944e-05, "loss": 0.0332, "step": 107700 }, { "epoch": 10.84, "learning_rate": 2.2962523958438416e-05, "loss": 0.0311, "step": 107750 }, { "epoch": 10.85, "learning_rate": 2.2949914254009887e-05, "loss": 0.0314, "step": 107800 }, { "epoch": 10.85, "learning_rate": 2.293730454958136e-05, "loss": 0.0283, "step": 107850 }, { "epoch": 10.86, "learning_rate": 2.29249470392414e-05, "loss": 0.0429, "step": 107900 }, { "epoch": 10.86, "learning_rate": 2.291233733481287e-05, "loss": 0.0302, "step": 107950 }, { "epoch": 10.87, "learning_rate": 2.2899727630384346e-05, "loss": 0.0323, "step": 108000 }, { "epoch": 10.87, "eval_cer": 0.06349977272585756, "eval_loss": 0.513024628162384, "eval_runtime": 1396.1047, "eval_samples_per_second": 5.841, "eval_steps_per_second": 0.731, "step": 108000 }, { "epoch": 10.87, "learning_rate": 2.2887117925955818e-05, "loss": 0.0316, "step": 108050 }, { "epoch": 10.88, "learning_rate": 2.287450822152729e-05, "loss": 0.0349, "step": 108100 }, { "epoch": 10.88, "learning_rate": 2.286189851709876e-05, "loss": 0.0274, "step": 108150 }, { "epoch": 10.89, "learning_rate": 2.2849288812670232e-05, "loss": 0.0266, "step": 108200 }, { "epoch": 10.89, "learning_rate": 2.2836679108241704e-05, "loss": 0.0344, "step": 108250 }, { "epoch": 10.9, "learning_rate": 2.2824069403813175e-05, "loss": 0.027, "step": 108300 }, { "epoch": 10.9, "learning_rate": 2.2811459699384647e-05, "loss": 0.0314, "step": 108350 }, { "epoch": 10.91, "learning_rate": 2.279884999495612e-05, "loss": 0.0323, "step": 108400 }, { "epoch": 10.91, "learning_rate": 2.278624029052759e-05, "loss": 0.0334, "step": 108450 }, { "epoch": 10.92, "learning_rate": 2.277363058609906e-05, "loss": 0.0338, "step": 108500 }, { "epoch": 10.92, "eval_cer": 0.06423763535264852, "eval_loss": 0.5162291526794434, "eval_runtime": 1340.5272, "eval_samples_per_second": 6.083, "eval_steps_per_second": 0.761, "step": 108500 }, { "epoch": 10.92, "learning_rate": 2.2761020881670533e-05, "loss": 0.0323, "step": 108550 }, { "epoch": 10.93, "learning_rate": 2.2748411177242007e-05, "loss": 0.0317, "step": 108600 }, { "epoch": 10.93, "learning_rate": 2.273580147281348e-05, "loss": 0.0316, "step": 108650 }, { "epoch": 10.94, "learning_rate": 2.272319176838495e-05, "loss": 0.0309, "step": 108700 }, { "epoch": 10.94, "learning_rate": 2.2710582063956422e-05, "loss": 0.0359, "step": 108750 }, { "epoch": 10.95, "learning_rate": 2.2697972359527893e-05, "loss": 0.0347, "step": 108800 }, { "epoch": 10.95, "learning_rate": 2.2685362655099365e-05, "loss": 0.038, "step": 108850 }, { "epoch": 10.96, "learning_rate": 2.2672752950670836e-05, "loss": 0.0332, "step": 108900 }, { "epoch": 10.96, "learning_rate": 2.266014324624231e-05, "loss": 0.0351, "step": 108950 }, { "epoch": 10.97, "learning_rate": 2.2647533541813783e-05, "loss": 0.0358, "step": 109000 }, { "epoch": 10.97, "eval_cer": 0.06348731934819021, "eval_loss": 0.5168716907501221, "eval_runtime": 1336.2307, "eval_samples_per_second": 6.103, "eval_steps_per_second": 0.763, "step": 109000 }, { "epoch": 10.97, "learning_rate": 2.263492383738525e-05, "loss": 0.0329, "step": 109050 }, { "epoch": 10.98, "learning_rate": 2.2622314132956722e-05, "loss": 0.0351, "step": 109100 }, { "epoch": 10.98, "learning_rate": 2.2609704428528197e-05, "loss": 0.0318, "step": 109150 }, { "epoch": 10.99, "learning_rate": 2.259709472409967e-05, "loss": 0.0303, "step": 109200 }, { "epoch": 10.99, "learning_rate": 2.258448501967114e-05, "loss": 0.0352, "step": 109250 }, { "epoch": 11.0, "learning_rate": 2.257187531524261e-05, "loss": 0.04, "step": 109300 }, { "epoch": 11.0, "learning_rate": 2.2559265610814083e-05, "loss": 0.0248, "step": 109350 }, { "epoch": 11.01, "learning_rate": 2.2546655906385555e-05, "loss": 0.0196, "step": 109400 }, { "epoch": 11.01, "learning_rate": 2.2534046201957026e-05, "loss": 0.0206, "step": 109450 }, { "epoch": 11.02, "learning_rate": 2.25214364975285e-05, "loss": 0.0221, "step": 109500 }, { "epoch": 11.02, "eval_cer": 0.06264360301122672, "eval_loss": 0.5192041993141174, "eval_runtime": 1332.1984, "eval_samples_per_second": 6.121, "eval_steps_per_second": 0.766, "step": 109500 }, { "epoch": 11.02, "learning_rate": 2.2508826793099972e-05, "loss": 0.0233, "step": 109550 }, { "epoch": 11.03, "learning_rate": 2.2496217088671444e-05, "loss": 0.0194, "step": 109600 }, { "epoch": 11.03, "learning_rate": 2.2483607384242912e-05, "loss": 0.021, "step": 109650 }, { "epoch": 11.04, "learning_rate": 2.2470997679814387e-05, "loss": 0.0186, "step": 109700 }, { "epoch": 11.04, "learning_rate": 2.245838797538586e-05, "loss": 0.0202, "step": 109750 }, { "epoch": 11.05, "learning_rate": 2.244577827095733e-05, "loss": 0.024, "step": 109800 }, { "epoch": 11.05, "learning_rate": 2.24331685665288e-05, "loss": 0.0235, "step": 109850 }, { "epoch": 11.06, "learning_rate": 2.2420558862100273e-05, "loss": 0.0204, "step": 109900 }, { "epoch": 11.06, "learning_rate": 2.2407949157671744e-05, "loss": 0.0253, "step": 109950 }, { "epoch": 11.07, "learning_rate": 2.2395339453243216e-05, "loss": 0.02, "step": 110000 }, { "epoch": 11.07, "eval_cer": 0.06325070517251041, "eval_loss": 0.5185605883598328, "eval_runtime": 1354.9428, "eval_samples_per_second": 6.019, "eval_steps_per_second": 0.753, "step": 110000 }, { "epoch": 11.07, "learning_rate": 2.2382729748814687e-05, "loss": 0.0202, "step": 110050 }, { "epoch": 11.08, "learning_rate": 2.2370120044386162e-05, "loss": 0.0195, "step": 110100 }, { "epoch": 11.08, "learning_rate": 2.2357510339957634e-05, "loss": 0.0238, "step": 110150 }, { "epoch": 11.09, "learning_rate": 2.2344900635529105e-05, "loss": 0.0208, "step": 110200 }, { "epoch": 11.09, "learning_rate": 2.2332290931100577e-05, "loss": 0.0237, "step": 110250 }, { "epoch": 11.1, "learning_rate": 2.2319681226672048e-05, "loss": 0.0214, "step": 110300 }, { "epoch": 11.1, "learning_rate": 2.230707152224352e-05, "loss": 0.0191, "step": 110350 }, { "epoch": 11.11, "learning_rate": 2.229446181781499e-05, "loss": 0.0207, "step": 110400 }, { "epoch": 11.11, "learning_rate": 2.2281852113386466e-05, "loss": 0.0201, "step": 110450 }, { "epoch": 11.12, "learning_rate": 2.2269242408957934e-05, "loss": 0.0221, "step": 110500 }, { "epoch": 11.12, "eval_cer": 0.06274634337698243, "eval_loss": 0.5232933163642883, "eval_runtime": 1478.2226, "eval_samples_per_second": 5.517, "eval_steps_per_second": 0.69, "step": 110500 }, { "epoch": 11.12, "learning_rate": 2.2256632704529405e-05, "loss": 0.0195, "step": 110550 }, { "epoch": 11.13, "learning_rate": 2.2244023000100877e-05, "loss": 0.0196, "step": 110600 }, { "epoch": 11.13, "learning_rate": 2.2231413295672352e-05, "loss": 0.0196, "step": 110650 }, { "epoch": 11.14, "learning_rate": 2.2218803591243823e-05, "loss": 0.023, "step": 110700 }, { "epoch": 11.14, "learning_rate": 2.2206193886815295e-05, "loss": 0.0225, "step": 110750 }, { "epoch": 11.15, "learning_rate": 2.2193584182386766e-05, "loss": 0.0249, "step": 110800 }, { "epoch": 11.15, "learning_rate": 2.2180974477958238e-05, "loss": 0.0219, "step": 110850 }, { "epoch": 11.16, "learning_rate": 2.216836477352971e-05, "loss": 0.0241, "step": 110900 }, { "epoch": 11.16, "learning_rate": 2.215575506910118e-05, "loss": 0.0193, "step": 110950 }, { "epoch": 11.17, "learning_rate": 2.2143145364672656e-05, "loss": 0.0252, "step": 111000 }, { "epoch": 11.17, "eval_cer": 0.06292069066432543, "eval_loss": 0.5196650624275208, "eval_runtime": 1654.4667, "eval_samples_per_second": 4.929, "eval_steps_per_second": 0.617, "step": 111000 }, { "epoch": 11.17, "learning_rate": 2.2130535660244127e-05, "loss": 0.0183, "step": 111050 }, { "epoch": 11.18, "learning_rate": 2.2117925955815595e-05, "loss": 0.0236, "step": 111100 }, { "epoch": 11.18, "learning_rate": 2.210556844547564e-05, "loss": 0.0257, "step": 111150 }, { "epoch": 11.19, "learning_rate": 2.209295874104711e-05, "loss": 0.0224, "step": 111200 }, { "epoch": 11.19, "learning_rate": 2.2080349036618583e-05, "loss": 0.0222, "step": 111250 }, { "epoch": 11.2, "learning_rate": 2.2067739332190054e-05, "loss": 0.0211, "step": 111300 }, { "epoch": 11.2, "learning_rate": 2.2055129627761526e-05, "loss": 0.0227, "step": 111350 }, { "epoch": 11.21, "learning_rate": 2.2042519923333e-05, "loss": 0.0197, "step": 111400 }, { "epoch": 11.21, "learning_rate": 2.202991021890447e-05, "loss": 0.0248, "step": 111450 }, { "epoch": 11.22, "learning_rate": 2.201730051447594e-05, "loss": 0.0256, "step": 111500 }, { "epoch": 11.22, "eval_cer": 0.0630514511298327, "eval_loss": 0.5244346857070923, "eval_runtime": 1648.9273, "eval_samples_per_second": 4.946, "eval_steps_per_second": 0.619, "step": 111500 }, { "epoch": 11.22, "learning_rate": 2.2004690810047412e-05, "loss": 0.0247, "step": 111550 }, { "epoch": 11.23, "learning_rate": 2.1992081105618887e-05, "loss": 0.0192, "step": 111600 }, { "epoch": 11.23, "learning_rate": 2.1979471401190358e-05, "loss": 0.0234, "step": 111650 }, { "epoch": 11.24, "learning_rate": 2.196686169676183e-05, "loss": 0.0196, "step": 111700 }, { "epoch": 11.24, "learning_rate": 2.19542519923333e-05, "loss": 0.0211, "step": 111750 }, { "epoch": 11.25, "learning_rate": 2.1941642287904773e-05, "loss": 0.0218, "step": 111800 }, { "epoch": 11.25, "learning_rate": 2.1929032583476244e-05, "loss": 0.0222, "step": 111850 }, { "epoch": 11.26, "learning_rate": 2.1916675073136285e-05, "loss": 0.0268, "step": 111900 }, { "epoch": 11.26, "learning_rate": 2.1904065368707757e-05, "loss": 0.0237, "step": 111950 }, { "epoch": 11.27, "learning_rate": 2.189145566427923e-05, "loss": 0.0217, "step": 112000 }, { "epoch": 11.27, "eval_cer": 0.06305767781866636, "eval_loss": 0.5219118595123291, "eval_runtime": 1635.4703, "eval_samples_per_second": 4.986, "eval_steps_per_second": 0.624, "step": 112000 }, { "epoch": 11.27, "learning_rate": 2.1878845959850703e-05, "loss": 0.0212, "step": 112050 }, { "epoch": 11.28, "learning_rate": 2.1866236255422175e-05, "loss": 0.016, "step": 112100 }, { "epoch": 11.28, "learning_rate": 2.1853626550993643e-05, "loss": 0.0197, "step": 112150 }, { "epoch": 11.29, "learning_rate": 2.1841016846565118e-05, "loss": 0.0209, "step": 112200 }, { "epoch": 11.3, "learning_rate": 2.182840714213659e-05, "loss": 0.0227, "step": 112250 }, { "epoch": 11.3, "learning_rate": 2.181579743770806e-05, "loss": 0.0186, "step": 112300 }, { "epoch": 11.31, "learning_rate": 2.1803187733279535e-05, "loss": 0.0195, "step": 112350 }, { "epoch": 11.31, "learning_rate": 2.1790578028851007e-05, "loss": 0.0186, "step": 112400 }, { "epoch": 11.32, "learning_rate": 2.1777968324422475e-05, "loss": 0.0209, "step": 112450 }, { "epoch": 11.32, "learning_rate": 2.1765358619993947e-05, "loss": 0.0215, "step": 112500 }, { "epoch": 11.32, "eval_cer": 0.06296116414174434, "eval_loss": 0.5222393870353699, "eval_runtime": 1350.9317, "eval_samples_per_second": 6.037, "eval_steps_per_second": 0.755, "step": 112500 }, { "epoch": 11.33, "learning_rate": 2.175274891556542e-05, "loss": 0.0213, "step": 112550 }, { "epoch": 11.33, "learning_rate": 2.1740139211136893e-05, "loss": 0.0217, "step": 112600 }, { "epoch": 11.34, "learning_rate": 2.1727529506708364e-05, "loss": 0.0216, "step": 112650 }, { "epoch": 11.34, "learning_rate": 2.1714919802279836e-05, "loss": 0.0243, "step": 112700 }, { "epoch": 11.35, "learning_rate": 2.1702310097851307e-05, "loss": 0.0231, "step": 112750 }, { "epoch": 11.35, "learning_rate": 2.168970039342278e-05, "loss": 0.0206, "step": 112800 }, { "epoch": 11.36, "learning_rate": 2.167709068899425e-05, "loss": 0.0212, "step": 112850 }, { "epoch": 11.36, "learning_rate": 2.1664480984565722e-05, "loss": 0.0281, "step": 112900 }, { "epoch": 11.37, "learning_rate": 2.1651871280137197e-05, "loss": 0.023, "step": 112950 }, { "epoch": 11.37, "learning_rate": 2.1639261575708665e-05, "loss": 0.0249, "step": 113000 }, { "epoch": 11.37, "eval_cer": 0.06244123562413215, "eval_loss": 0.5205076932907104, "eval_runtime": 1350.6359, "eval_samples_per_second": 6.038, "eval_steps_per_second": 0.755, "step": 113000 }, { "epoch": 11.38, "learning_rate": 2.1626651871280136e-05, "loss": 0.0208, "step": 113050 }, { "epoch": 11.38, "learning_rate": 2.161404216685161e-05, "loss": 0.0199, "step": 113100 }, { "epoch": 11.39, "learning_rate": 2.1601432462423083e-05, "loss": 0.0217, "step": 113150 }, { "epoch": 11.39, "learning_rate": 2.1588822757994554e-05, "loss": 0.0216, "step": 113200 }, { "epoch": 11.4, "learning_rate": 2.1576213053566026e-05, "loss": 0.0244, "step": 113250 }, { "epoch": 11.4, "learning_rate": 2.1563603349137497e-05, "loss": 0.0215, "step": 113300 }, { "epoch": 11.41, "learning_rate": 2.155099364470897e-05, "loss": 0.022, "step": 113350 }, { "epoch": 11.41, "learning_rate": 2.153838394028044e-05, "loss": 0.0234, "step": 113400 }, { "epoch": 11.42, "learning_rate": 2.152577423585191e-05, "loss": 0.0196, "step": 113450 }, { "epoch": 11.42, "learning_rate": 2.1513164531423386e-05, "loss": 0.0232, "step": 113500 }, { "epoch": 11.42, "eval_cer": 0.0629020105978244, "eval_loss": 0.5262394547462463, "eval_runtime": 1343.7609, "eval_samples_per_second": 6.069, "eval_steps_per_second": 0.759, "step": 113500 }, { "epoch": 11.43, "learning_rate": 2.1500554826994858e-05, "loss": 0.0217, "step": 113550 }, { "epoch": 11.43, "learning_rate": 2.1487945122566326e-05, "loss": 0.0217, "step": 113600 }, { "epoch": 11.44, "learning_rate": 2.1475335418137797e-05, "loss": 0.0228, "step": 113650 }, { "epoch": 11.44, "learning_rate": 2.1462725713709272e-05, "loss": 0.0247, "step": 113700 }, { "epoch": 11.45, "learning_rate": 2.1450116009280744e-05, "loss": 0.0215, "step": 113750 }, { "epoch": 11.45, "learning_rate": 2.1437506304852215e-05, "loss": 0.02, "step": 113800 }, { "epoch": 11.46, "learning_rate": 2.1424896600423687e-05, "loss": 0.0205, "step": 113850 }, { "epoch": 11.46, "learning_rate": 2.1412286895995158e-05, "loss": 0.0231, "step": 113900 }, { "epoch": 11.47, "learning_rate": 2.139967719156663e-05, "loss": 0.0256, "step": 113950 }, { "epoch": 11.47, "learning_rate": 2.13870674871381e-05, "loss": 0.0229, "step": 114000 }, { "epoch": 11.47, "eval_cer": 0.06311371801816948, "eval_loss": 0.5250869393348694, "eval_runtime": 1643.2324, "eval_samples_per_second": 4.963, "eval_steps_per_second": 0.621, "step": 114000 }, { "epoch": 11.48, "learning_rate": 2.1374457782709576e-05, "loss": 0.0271, "step": 114050 }, { "epoch": 11.48, "learning_rate": 2.1361848078281048e-05, "loss": 0.0234, "step": 114100 }, { "epoch": 11.49, "learning_rate": 2.134923837385252e-05, "loss": 0.025, "step": 114150 }, { "epoch": 11.49, "learning_rate": 2.1336628669423987e-05, "loss": 0.0242, "step": 114200 }, { "epoch": 11.5, "learning_rate": 2.1324018964995462e-05, "loss": 0.0229, "step": 114250 }, { "epoch": 11.5, "learning_rate": 2.1311409260566933e-05, "loss": 0.0214, "step": 114300 }, { "epoch": 11.51, "learning_rate": 2.1298799556138405e-05, "loss": 0.0211, "step": 114350 }, { "epoch": 11.51, "learning_rate": 2.1286189851709876e-05, "loss": 0.0209, "step": 114400 }, { "epoch": 11.52, "learning_rate": 2.1273580147281348e-05, "loss": 0.0253, "step": 114450 }, { "epoch": 11.52, "learning_rate": 2.126097044285282e-05, "loss": 0.0224, "step": 114500 }, { "epoch": 11.52, "eval_cer": 0.06320711835067466, "eval_loss": 0.5250407457351685, "eval_runtime": 1652.0069, "eval_samples_per_second": 4.936, "eval_steps_per_second": 0.617, "step": 114500 }, { "epoch": 11.53, "learning_rate": 2.124836073842429e-05, "loss": 0.0211, "step": 114550 }, { "epoch": 11.53, "learning_rate": 2.1235751033995766e-05, "loss": 0.0229, "step": 114600 }, { "epoch": 11.54, "learning_rate": 2.1223141329567237e-05, "loss": 0.0256, "step": 114650 }, { "epoch": 11.54, "learning_rate": 2.121053162513871e-05, "loss": 0.0267, "step": 114700 }, { "epoch": 11.55, "learning_rate": 2.119792192071018e-05, "loss": 0.0244, "step": 114750 }, { "epoch": 11.55, "learning_rate": 2.118531221628165e-05, "loss": 0.0213, "step": 114800 }, { "epoch": 11.56, "learning_rate": 2.1172702511853123e-05, "loss": 0.0292, "step": 114850 }, { "epoch": 11.56, "learning_rate": 2.1160092807424595e-05, "loss": 0.021, "step": 114900 }, { "epoch": 11.57, "learning_rate": 2.1147483102996066e-05, "loss": 0.0259, "step": 114950 }, { "epoch": 11.57, "learning_rate": 2.113487339856754e-05, "loss": 0.0247, "step": 115000 }, { "epoch": 11.57, "eval_cer": 0.06307635788516741, "eval_loss": 0.5251887440681458, "eval_runtime": 1533.7763, "eval_samples_per_second": 5.317, "eval_steps_per_second": 0.665, "step": 115000 }, { "epoch": 11.58, "learning_rate": 2.112226369413901e-05, "loss": 0.0249, "step": 115050 }, { "epoch": 11.58, "learning_rate": 2.110965398971048e-05, "loss": 0.0225, "step": 115100 }, { "epoch": 11.59, "learning_rate": 2.1097044285281952e-05, "loss": 0.0267, "step": 115150 }, { "epoch": 11.59, "learning_rate": 2.1084434580853427e-05, "loss": 0.0413, "step": 115200 }, { "epoch": 11.6, "learning_rate": 2.10718248764249e-05, "loss": 0.0213, "step": 115250 }, { "epoch": 11.6, "learning_rate": 2.105921517199637e-05, "loss": 0.0254, "step": 115300 }, { "epoch": 11.61, "learning_rate": 2.104660546756784e-05, "loss": 0.0252, "step": 115350 }, { "epoch": 11.61, "learning_rate": 2.1033995763139313e-05, "loss": 0.0251, "step": 115400 }, { "epoch": 11.62, "learning_rate": 2.1021386058710784e-05, "loss": 0.0254, "step": 115450 }, { "epoch": 11.62, "learning_rate": 2.1008776354282256e-05, "loss": 0.0208, "step": 115500 }, { "epoch": 11.62, "eval_cer": 0.06255954271197206, "eval_loss": 0.5247114300727844, "eval_runtime": 1639.1923, "eval_samples_per_second": 4.975, "eval_steps_per_second": 0.622, "step": 115500 }, { "epoch": 11.63, "learning_rate": 2.099616664985373e-05, "loss": 0.0214, "step": 115550 }, { "epoch": 11.63, "learning_rate": 2.0983556945425202e-05, "loss": 0.0254, "step": 115600 }, { "epoch": 11.64, "learning_rate": 2.097094724099667e-05, "loss": 0.023, "step": 115650 }, { "epoch": 11.64, "learning_rate": 2.0958337536568142e-05, "loss": 0.022, "step": 115700 }, { "epoch": 11.65, "learning_rate": 2.0945727832139617e-05, "loss": 0.0253, "step": 115750 }, { "epoch": 11.65, "learning_rate": 2.0933118127711088e-05, "loss": 0.0231, "step": 115800 }, { "epoch": 11.66, "learning_rate": 2.092050842328256e-05, "loss": 0.0233, "step": 115850 }, { "epoch": 11.66, "learning_rate": 2.090789871885403e-05, "loss": 0.0237, "step": 115900 }, { "epoch": 11.67, "learning_rate": 2.0895289014425503e-05, "loss": 0.0249, "step": 115950 }, { "epoch": 11.67, "learning_rate": 2.0882679309996974e-05, "loss": 0.0222, "step": 116000 }, { "epoch": 11.67, "eval_cer": 0.06335344553826612, "eval_loss": 0.5233452916145325, "eval_runtime": 1606.1203, "eval_samples_per_second": 5.077, "eval_steps_per_second": 0.635, "step": 116000 }, { "epoch": 11.68, "learning_rate": 2.0870069605568446e-05, "loss": 0.0277, "step": 116050 }, { "epoch": 11.68, "learning_rate": 2.085745990113992e-05, "loss": 0.0224, "step": 116100 }, { "epoch": 11.69, "learning_rate": 2.0844850196711392e-05, "loss": 0.0242, "step": 116150 }, { "epoch": 11.69, "learning_rate": 2.0832240492282863e-05, "loss": 0.0217, "step": 116200 }, { "epoch": 11.7, "learning_rate": 2.081963078785433e-05, "loss": 0.0239, "step": 116250 }, { "epoch": 11.7, "learning_rate": 2.0807021083425806e-05, "loss": 0.0214, "step": 116300 }, { "epoch": 11.71, "learning_rate": 2.0794411378997278e-05, "loss": 0.022, "step": 116350 }, { "epoch": 11.71, "learning_rate": 2.078180167456875e-05, "loss": 0.0234, "step": 116400 }, { "epoch": 11.72, "learning_rate": 2.076919197014022e-05, "loss": 0.0241, "step": 116450 }, { "epoch": 11.72, "learning_rate": 2.0756582265711692e-05, "loss": 0.0232, "step": 116500 }, { "epoch": 11.72, "eval_cer": 0.063297405338763, "eval_loss": 0.5278661847114563, "eval_runtime": 1580.2617, "eval_samples_per_second": 5.161, "eval_steps_per_second": 0.645, "step": 116500 }, { "epoch": 11.73, "learning_rate": 2.0743972561283164e-05, "loss": 0.0255, "step": 116550 }, { "epoch": 11.73, "learning_rate": 2.0731362856854635e-05, "loss": 0.0239, "step": 116600 }, { "epoch": 11.74, "learning_rate": 2.0718753152426107e-05, "loss": 0.025, "step": 116650 }, { "epoch": 11.74, "learning_rate": 2.070614344799758e-05, "loss": 0.0232, "step": 116700 }, { "epoch": 11.75, "learning_rate": 2.0693533743569053e-05, "loss": 0.0233, "step": 116750 }, { "epoch": 11.75, "learning_rate": 2.0680924039140525e-05, "loss": 0.0227, "step": 116800 }, { "epoch": 11.76, "learning_rate": 2.0668314334711993e-05, "loss": 0.0207, "step": 116850 }, { "epoch": 11.76, "learning_rate": 2.0655704630283468e-05, "loss": 0.0222, "step": 116900 }, { "epoch": 11.77, "learning_rate": 2.064309492585494e-05, "loss": 0.0259, "step": 116950 }, { "epoch": 11.77, "learning_rate": 2.063048522142641e-05, "loss": 0.024, "step": 117000 }, { "epoch": 11.77, "eval_cer": 0.06397922776605085, "eval_loss": 0.5233616232872009, "eval_runtime": 1307.5213, "eval_samples_per_second": 6.237, "eval_steps_per_second": 0.78, "step": 117000 }, { "epoch": 11.78, "learning_rate": 2.0617875516997885e-05, "loss": 0.0214, "step": 117050 }, { "epoch": 11.78, "learning_rate": 2.0605265812569353e-05, "loss": 0.0206, "step": 117100 }, { "epoch": 11.79, "learning_rate": 2.0592656108140825e-05, "loss": 0.0224, "step": 117150 }, { "epoch": 11.79, "learning_rate": 2.0580046403712296e-05, "loss": 0.024, "step": 117200 }, { "epoch": 11.8, "learning_rate": 2.056743669928377e-05, "loss": 0.0232, "step": 117250 }, { "epoch": 11.8, "learning_rate": 2.0554826994855243e-05, "loss": 0.022, "step": 117300 }, { "epoch": 11.81, "learning_rate": 2.0542217290426714e-05, "loss": 0.0226, "step": 117350 }, { "epoch": 11.81, "learning_rate": 2.0529607585998182e-05, "loss": 0.0246, "step": 117400 }, { "epoch": 11.82, "learning_rate": 2.0516997881569657e-05, "loss": 0.0216, "step": 117450 }, { "epoch": 11.82, "learning_rate": 2.050438817714113e-05, "loss": 0.0221, "step": 117500 }, { "epoch": 11.82, "eval_cer": 0.06337835229360084, "eval_loss": 0.5259285569190979, "eval_runtime": 1298.8721, "eval_samples_per_second": 6.279, "eval_steps_per_second": 0.785, "step": 117500 }, { "epoch": 11.83, "learning_rate": 2.04917784727126e-05, "loss": 0.0248, "step": 117550 }, { "epoch": 11.83, "learning_rate": 2.0479168768284072e-05, "loss": 0.0223, "step": 117600 }, { "epoch": 11.84, "learning_rate": 2.0466559063855547e-05, "loss": 0.0264, "step": 117650 }, { "epoch": 11.84, "learning_rate": 2.0453949359427015e-05, "loss": 0.027, "step": 117700 }, { "epoch": 11.85, "learning_rate": 2.0441339654998486e-05, "loss": 0.026, "step": 117750 }, { "epoch": 11.85, "learning_rate": 2.042872995056996e-05, "loss": 0.0278, "step": 117800 }, { "epoch": 11.86, "learning_rate": 2.0416120246141433e-05, "loss": 0.0236, "step": 117850 }, { "epoch": 11.86, "learning_rate": 2.0403510541712904e-05, "loss": 0.0268, "step": 117900 }, { "epoch": 11.87, "learning_rate": 2.0390900837284375e-05, "loss": 0.0235, "step": 117950 }, { "epoch": 11.87, "learning_rate": 2.0378291132855847e-05, "loss": 0.0243, "step": 118000 }, { "epoch": 11.87, "eval_cer": 0.06274011668814874, "eval_loss": 0.5240441560745239, "eval_runtime": 1311.4796, "eval_samples_per_second": 6.218, "eval_steps_per_second": 0.778, "step": 118000 }, { "epoch": 11.88, "learning_rate": 2.036568142842732e-05, "loss": 0.0219, "step": 118050 }, { "epoch": 11.88, "learning_rate": 2.035307172399879e-05, "loss": 0.021, "step": 118100 }, { "epoch": 11.89, "learning_rate": 2.034046201957026e-05, "loss": 0.0261, "step": 118150 }, { "epoch": 11.89, "learning_rate": 2.0327852315141736e-05, "loss": 0.0254, "step": 118200 }, { "epoch": 11.9, "learning_rate": 2.0315242610713204e-05, "loss": 0.0263, "step": 118250 }, { "epoch": 11.9, "learning_rate": 2.0302632906284676e-05, "loss": 0.0282, "step": 118300 }, { "epoch": 11.91, "learning_rate": 2.0290023201856147e-05, "loss": 0.0223, "step": 118350 }, { "epoch": 11.91, "learning_rate": 2.0277413497427622e-05, "loss": 0.028, "step": 118400 }, { "epoch": 11.92, "learning_rate": 2.0264803792999094e-05, "loss": 0.0238, "step": 118450 }, { "epoch": 11.92, "learning_rate": 2.0252194088570565e-05, "loss": 0.0194, "step": 118500 }, { "epoch": 11.92, "eval_cer": 0.0628864438757402, "eval_loss": 0.5283812880516052, "eval_runtime": 1666.8387, "eval_samples_per_second": 4.892, "eval_steps_per_second": 0.612, "step": 118500 }, { "epoch": 11.93, "learning_rate": 2.0239584384142037e-05, "loss": 0.0234, "step": 118550 }, { "epoch": 11.93, "learning_rate": 2.0226974679713508e-05, "loss": 0.0228, "step": 118600 }, { "epoch": 11.94, "learning_rate": 2.021436497528498e-05, "loss": 0.0206, "step": 118650 }, { "epoch": 11.94, "learning_rate": 2.020175527085645e-05, "loss": 0.0223, "step": 118700 }, { "epoch": 11.95, "learning_rate": 2.0189145566427926e-05, "loss": 0.0245, "step": 118750 }, { "epoch": 11.95, "learning_rate": 2.0176535861999397e-05, "loss": 0.0221, "step": 118800 }, { "epoch": 11.96, "learning_rate": 2.0163926157570866e-05, "loss": 0.0231, "step": 118850 }, { "epoch": 11.96, "learning_rate": 2.0151316453142337e-05, "loss": 0.0262, "step": 118900 }, { "epoch": 11.97, "learning_rate": 2.0138706748713812e-05, "loss": 0.024, "step": 118950 }, { "epoch": 11.97, "learning_rate": 2.0126097044285283e-05, "loss": 0.0234, "step": 119000 }, { "epoch": 11.97, "eval_cer": 0.06315107815117155, "eval_loss": 0.5276172161102295, "eval_runtime": 1662.9563, "eval_samples_per_second": 4.904, "eval_steps_per_second": 0.613, "step": 119000 }, { "epoch": 11.98, "learning_rate": 2.0113487339856755e-05, "loss": 0.0224, "step": 119050 }, { "epoch": 11.98, "learning_rate": 2.0100877635428226e-05, "loss": 0.0218, "step": 119100 }, { "epoch": 11.99, "learning_rate": 2.0088267930999698e-05, "loss": 0.0224, "step": 119150 }, { "epoch": 11.99, "learning_rate": 2.007565822657117e-05, "loss": 0.0248, "step": 119200 }, { "epoch": 12.0, "learning_rate": 2.006304852214264e-05, "loss": 0.028, "step": 119250 }, { "epoch": 12.0, "learning_rate": 2.0050438817714116e-05, "loss": 0.0183, "step": 119300 }, { "epoch": 12.01, "learning_rate": 2.0037829113285587e-05, "loss": 0.014, "step": 119350 }, { "epoch": 12.01, "learning_rate": 2.002521940885706e-05, "loss": 0.0147, "step": 119400 }, { "epoch": 12.02, "learning_rate": 2.0012609704428527e-05, "loss": 0.0167, "step": 119450 }, { "epoch": 12.02, "learning_rate": 2e-05, "loss": 0.0132, "step": 119500 }, { "epoch": 12.02, "eval_cer": 0.06200848075019147, "eval_loss": 0.5311650037765503, "eval_runtime": 1644.0071, "eval_samples_per_second": 4.96, "eval_steps_per_second": 0.62, "step": 119500 }, { "epoch": 12.03, "learning_rate": 1.9987390295571473e-05, "loss": 0.0177, "step": 119550 }, { "epoch": 12.03, "learning_rate": 1.9974780591142945e-05, "loss": 0.0133, "step": 119600 }, { "epoch": 12.04, "learning_rate": 1.9962170886714416e-05, "loss": 0.016, "step": 119650 }, { "epoch": 12.04, "learning_rate": 1.9949561182285888e-05, "loss": 0.015, "step": 119700 }, { "epoch": 12.05, "learning_rate": 1.993695147785736e-05, "loss": 0.0147, "step": 119750 }, { "epoch": 12.05, "learning_rate": 1.992434177342883e-05, "loss": 0.0135, "step": 119800 }, { "epoch": 12.06, "learning_rate": 1.9911732069000302e-05, "loss": 0.0172, "step": 119850 }, { "epoch": 12.06, "learning_rate": 1.9899122364571777e-05, "loss": 0.0156, "step": 119900 }, { "epoch": 12.07, "learning_rate": 1.988651266014325e-05, "loss": 0.0139, "step": 119950 }, { "epoch": 12.07, "learning_rate": 1.987415514980329e-05, "loss": 0.016, "step": 120000 }, { "epoch": 12.07, "eval_cer": 0.06191819376210313, "eval_loss": 0.5298161506652832, "eval_runtime": 1626.7108, "eval_samples_per_second": 5.013, "eval_steps_per_second": 0.627, "step": 120000 }, { "epoch": 12.08, "learning_rate": 1.986154544537476e-05, "loss": 0.0117, "step": 120050 }, { "epoch": 12.08, "learning_rate": 1.9848935740946233e-05, "loss": 0.0143, "step": 120100 }, { "epoch": 12.09, "learning_rate": 1.9836326036517704e-05, "loss": 0.0254, "step": 120150 }, { "epoch": 12.09, "learning_rate": 1.9823716332089176e-05, "loss": 0.0154, "step": 120200 }, { "epoch": 12.1, "learning_rate": 1.981110662766065e-05, "loss": 0.0177, "step": 120250 }, { "epoch": 12.11, "learning_rate": 1.9798496923232122e-05, "loss": 0.0171, "step": 120300 }, { "epoch": 12.11, "learning_rate": 1.9785887218803593e-05, "loss": 0.0135, "step": 120350 }, { "epoch": 12.12, "learning_rate": 1.977327751437506e-05, "loss": 0.0139, "step": 120400 }, { "epoch": 12.12, "learning_rate": 1.9760667809946536e-05, "loss": 0.0159, "step": 120450 }, { "epoch": 12.13, "learning_rate": 1.9748058105518008e-05, "loss": 0.018, "step": 120500 }, { "epoch": 12.13, "eval_cer": 0.062226414859370235, "eval_loss": 0.5302690267562866, "eval_runtime": 1318.7043, "eval_samples_per_second": 6.184, "eval_steps_per_second": 0.773, "step": 120500 }, { "epoch": 12.13, "learning_rate": 1.973544840108948e-05, "loss": 0.0155, "step": 120550 }, { "epoch": 12.14, "learning_rate": 1.972283869666095e-05, "loss": 0.0132, "step": 120600 }, { "epoch": 12.14, "learning_rate": 1.9710228992232422e-05, "loss": 0.014, "step": 120650 }, { "epoch": 12.15, "learning_rate": 1.9697619287803894e-05, "loss": 0.0197, "step": 120700 }, { "epoch": 12.15, "learning_rate": 1.9685009583375365e-05, "loss": 0.0145, "step": 120750 }, { "epoch": 12.16, "learning_rate": 1.9672399878946837e-05, "loss": 0.0169, "step": 120800 }, { "epoch": 12.16, "learning_rate": 1.965979017451831e-05, "loss": 0.0146, "step": 120850 }, { "epoch": 12.17, "learning_rate": 1.9647180470089783e-05, "loss": 0.0146, "step": 120900 }, { "epoch": 12.17, "learning_rate": 1.9634570765661255e-05, "loss": 0.0142, "step": 120950 }, { "epoch": 12.18, "learning_rate": 1.9621961061232723e-05, "loss": 0.0127, "step": 121000 }, { "epoch": 12.18, "eval_cer": 0.062226414859370235, "eval_loss": 0.5325795412063599, "eval_runtime": 1648.4485, "eval_samples_per_second": 4.947, "eval_steps_per_second": 0.619, "step": 121000 }, { "epoch": 12.18, "learning_rate": 1.9609351356804198e-05, "loss": 0.0157, "step": 121050 }, { "epoch": 12.19, "learning_rate": 1.959674165237567e-05, "loss": 0.0198, "step": 121100 }, { "epoch": 12.19, "learning_rate": 1.958413194794714e-05, "loss": 0.0145, "step": 121150 }, { "epoch": 12.2, "learning_rate": 1.9571522243518615e-05, "loss": 0.0146, "step": 121200 }, { "epoch": 12.2, "learning_rate": 1.9558912539090084e-05, "loss": 0.0139, "step": 121250 }, { "epoch": 12.21, "learning_rate": 1.9546302834661555e-05, "loss": 0.0184, "step": 121300 }, { "epoch": 12.21, "learning_rate": 1.9533693130233027e-05, "loss": 0.0156, "step": 121350 }, { "epoch": 12.22, "learning_rate": 1.95210834258045e-05, "loss": 0.0155, "step": 121400 }, { "epoch": 12.22, "learning_rate": 1.9508473721375973e-05, "loss": 0.0185, "step": 121450 }, { "epoch": 12.23, "learning_rate": 1.9495864016947444e-05, "loss": 0.0153, "step": 121500 }, { "epoch": 12.23, "eval_cer": 0.06184970018493266, "eval_loss": 0.5320751667022705, "eval_runtime": 1651.1166, "eval_samples_per_second": 4.939, "eval_steps_per_second": 0.618, "step": 121500 }, { "epoch": 12.23, "learning_rate": 1.9483254312518916e-05, "loss": 0.0155, "step": 121550 }, { "epoch": 12.24, "learning_rate": 1.9470644608090387e-05, "loss": 0.0161, "step": 121600 }, { "epoch": 12.24, "learning_rate": 1.945803490366186e-05, "loss": 0.0141, "step": 121650 }, { "epoch": 12.25, "learning_rate": 1.944542519923333e-05, "loss": 0.0149, "step": 121700 }, { "epoch": 12.25, "learning_rate": 1.9432815494804802e-05, "loss": 0.0148, "step": 121750 }, { "epoch": 12.26, "learning_rate": 1.9420205790376277e-05, "loss": 0.0165, "step": 121800 }, { "epoch": 12.26, "learning_rate": 1.9407596085947745e-05, "loss": 0.0156, "step": 121850 }, { "epoch": 12.27, "learning_rate": 1.9394986381519216e-05, "loss": 0.0158, "step": 121900 }, { "epoch": 12.27, "learning_rate": 1.938237667709069e-05, "loss": 0.0148, "step": 121950 }, { "epoch": 12.28, "learning_rate": 1.9369766972662163e-05, "loss": 0.0127, "step": 122000 }, { "epoch": 12.28, "eval_cer": 0.0628148369541529, "eval_loss": 0.533798098564148, "eval_runtime": 1640.2456, "eval_samples_per_second": 4.972, "eval_steps_per_second": 0.622, "step": 122000 }, { "epoch": 12.28, "learning_rate": 1.9357157268233634e-05, "loss": 0.0143, "step": 122050 }, { "epoch": 12.29, "learning_rate": 1.9344547563805106e-05, "loss": 0.0148, "step": 122100 }, { "epoch": 12.29, "learning_rate": 1.933219005346515e-05, "loss": 0.0149, "step": 122150 }, { "epoch": 12.3, "learning_rate": 1.9319580349036622e-05, "loss": 0.0154, "step": 122200 }, { "epoch": 12.3, "learning_rate": 1.930697064460809e-05, "loss": 0.0142, "step": 122250 }, { "epoch": 12.31, "learning_rate": 1.929436094017956e-05, "loss": 0.015, "step": 122300 }, { "epoch": 12.31, "learning_rate": 1.9281751235751036e-05, "loss": 0.0148, "step": 122350 }, { "epoch": 12.32, "learning_rate": 1.9269141531322508e-05, "loss": 0.0162, "step": 122400 }, { "epoch": 12.32, "learning_rate": 1.925653182689398e-05, "loss": 0.0156, "step": 122450 }, { "epoch": 12.33, "learning_rate": 1.924392212246545e-05, "loss": 0.0169, "step": 122500 }, { "epoch": 12.33, "eval_cer": 0.061634879420170735, "eval_loss": 0.5355355739593506, "eval_runtime": 1530.1242, "eval_samples_per_second": 5.33, "eval_steps_per_second": 0.667, "step": 122500 }, { "epoch": 12.33, "learning_rate": 1.9231312418036922e-05, "loss": 0.0182, "step": 122550 }, { "epoch": 12.34, "learning_rate": 1.9218702713608394e-05, "loss": 0.0172, "step": 122600 }, { "epoch": 12.34, "learning_rate": 1.9206093009179865e-05, "loss": 0.0144, "step": 122650 }, { "epoch": 12.35, "learning_rate": 1.9193483304751337e-05, "loss": 0.0231, "step": 122700 }, { "epoch": 12.35, "learning_rate": 1.918087360032281e-05, "loss": 0.0129, "step": 122750 }, { "epoch": 12.36, "learning_rate": 1.916826389589428e-05, "loss": 0.0152, "step": 122800 }, { "epoch": 12.36, "learning_rate": 1.915565419146575e-05, "loss": 0.0199, "step": 122850 }, { "epoch": 12.37, "learning_rate": 1.9143044487037226e-05, "loss": 0.0156, "step": 122900 }, { "epoch": 12.37, "learning_rate": 1.9130434782608697e-05, "loss": 0.017, "step": 122950 }, { "epoch": 12.38, "learning_rate": 1.911782507818017e-05, "loss": 0.0154, "step": 123000 }, { "epoch": 12.38, "eval_cer": 0.0627370033437319, "eval_loss": 0.5299083590507507, "eval_runtime": 1556.576, "eval_samples_per_second": 5.239, "eval_steps_per_second": 0.655, "step": 123000 }, { "epoch": 12.38, "learning_rate": 1.910521537375164e-05, "loss": 0.0165, "step": 123050 }, { "epoch": 12.39, "learning_rate": 1.9092605669323112e-05, "loss": 0.0154, "step": 123100 }, { "epoch": 12.39, "learning_rate": 1.9079995964894583e-05, "loss": 0.0171, "step": 123150 }, { "epoch": 12.4, "learning_rate": 1.9067386260466055e-05, "loss": 0.0152, "step": 123200 }, { "epoch": 12.4, "learning_rate": 1.9054776556037526e-05, "loss": 0.0165, "step": 123250 }, { "epoch": 12.41, "learning_rate": 1.9042166851609e-05, "loss": 0.0168, "step": 123300 }, { "epoch": 12.41, "learning_rate": 1.9029557147180473e-05, "loss": 0.0185, "step": 123350 }, { "epoch": 12.42, "learning_rate": 1.901694744275194e-05, "loss": 0.0194, "step": 123400 }, { "epoch": 12.42, "learning_rate": 1.9004337738323412e-05, "loss": 0.0163, "step": 123450 }, { "epoch": 12.43, "learning_rate": 1.8991728033894887e-05, "loss": 0.0164, "step": 123500 }, { "epoch": 12.43, "eval_cer": 0.06233849525837645, "eval_loss": 0.538597047328949, "eval_runtime": 1338.5207, "eval_samples_per_second": 6.093, "eval_steps_per_second": 0.762, "step": 123500 }, { "epoch": 12.43, "learning_rate": 1.897911832946636e-05, "loss": 0.0169, "step": 123550 }, { "epoch": 12.44, "learning_rate": 1.896650862503783e-05, "loss": 0.0159, "step": 123600 }, { "epoch": 12.44, "learning_rate": 1.89538989206093e-05, "loss": 0.0134, "step": 123650 }, { "epoch": 12.45, "learning_rate": 1.8941289216180773e-05, "loss": 0.0178, "step": 123700 }, { "epoch": 12.45, "learning_rate": 1.8928679511752244e-05, "loss": 0.0159, "step": 123750 }, { "epoch": 12.46, "learning_rate": 1.8916069807323716e-05, "loss": 0.0152, "step": 123800 }, { "epoch": 12.46, "learning_rate": 1.890346010289519e-05, "loss": 0.0165, "step": 123850 }, { "epoch": 12.47, "learning_rate": 1.8890850398466662e-05, "loss": 0.0181, "step": 123900 }, { "epoch": 12.47, "learning_rate": 1.8878240694038134e-05, "loss": 0.0177, "step": 123950 }, { "epoch": 12.48, "learning_rate": 1.8865630989609602e-05, "loss": 0.0158, "step": 124000 }, { "epoch": 12.48, "eval_cer": 0.062285568403290185, "eval_loss": 0.5359772443771362, "eval_runtime": 1308.5707, "eval_samples_per_second": 6.232, "eval_steps_per_second": 0.779, "step": 124000 }, { "epoch": 12.48, "learning_rate": 1.8853021285181077e-05, "loss": 0.0173, "step": 124050 }, { "epoch": 12.49, "learning_rate": 1.8840411580752548e-05, "loss": 0.0144, "step": 124100 }, { "epoch": 12.49, "learning_rate": 1.882780187632402e-05, "loss": 0.0186, "step": 124150 }, { "epoch": 12.5, "learning_rate": 1.881519217189549e-05, "loss": 0.0191, "step": 124200 }, { "epoch": 12.5, "learning_rate": 1.8802582467466963e-05, "loss": 0.0155, "step": 124250 }, { "epoch": 12.51, "learning_rate": 1.8789972763038434e-05, "loss": 0.0171, "step": 124300 }, { "epoch": 12.51, "learning_rate": 1.8777363058609906e-05, "loss": 0.0168, "step": 124350 }, { "epoch": 12.52, "learning_rate": 1.876475335418138e-05, "loss": 0.0165, "step": 124400 }, { "epoch": 12.52, "learning_rate": 1.8752143649752852e-05, "loss": 0.0174, "step": 124450 }, { "epoch": 12.53, "learning_rate": 1.8739786139412893e-05, "loss": 0.018, "step": 124500 }, { "epoch": 12.53, "eval_cer": 0.062192168070785, "eval_loss": 0.5344801545143127, "eval_runtime": 1308.1584, "eval_samples_per_second": 6.234, "eval_steps_per_second": 0.78, "step": 124500 }, { "epoch": 12.53, "learning_rate": 1.8727176434984365e-05, "loss": 0.0192, "step": 124550 }, { "epoch": 12.54, "learning_rate": 1.8714566730555836e-05, "loss": 0.0177, "step": 124600 }, { "epoch": 12.54, "learning_rate": 1.8701957026127308e-05, "loss": 0.016, "step": 124650 }, { "epoch": 12.55, "learning_rate": 1.868934732169878e-05, "loss": 0.015, "step": 124700 }, { "epoch": 12.55, "learning_rate": 1.867673761727025e-05, "loss": 0.0157, "step": 124750 }, { "epoch": 12.56, "learning_rate": 1.8664127912841726e-05, "loss": 0.0156, "step": 124800 }, { "epoch": 12.56, "learning_rate": 1.8651518208413197e-05, "loss": 0.0159, "step": 124850 }, { "epoch": 12.57, "learning_rate": 1.863890850398467e-05, "loss": 0.0173, "step": 124900 }, { "epoch": 12.57, "learning_rate": 1.8626298799556137e-05, "loss": 0.0168, "step": 124950 }, { "epoch": 12.58, "learning_rate": 1.861368909512761e-05, "loss": 0.0158, "step": 125000 }, { "epoch": 12.58, "eval_cer": 0.06247548241271739, "eval_loss": 0.5364932417869568, "eval_runtime": 1433.4067, "eval_samples_per_second": 5.689, "eval_steps_per_second": 0.712, "step": 125000 }, { "epoch": 12.58, "learning_rate": 1.8601079390699083e-05, "loss": 0.0168, "step": 125050 }, { "epoch": 12.59, "learning_rate": 1.8588469686270555e-05, "loss": 0.0153, "step": 125100 }, { "epoch": 12.59, "learning_rate": 1.8575859981842026e-05, "loss": 0.0159, "step": 125150 }, { "epoch": 12.6, "learning_rate": 1.85632502774135e-05, "loss": 0.0183, "step": 125200 }, { "epoch": 12.6, "learning_rate": 1.855064057298497e-05, "loss": 0.0161, "step": 125250 }, { "epoch": 12.61, "learning_rate": 1.853803086855644e-05, "loss": 0.018, "step": 125300 }, { "epoch": 12.61, "learning_rate": 1.8525421164127912e-05, "loss": 0.0155, "step": 125350 }, { "epoch": 12.62, "learning_rate": 1.8512811459699387e-05, "loss": 0.0177, "step": 125400 }, { "epoch": 12.62, "learning_rate": 1.8500201755270858e-05, "loss": 0.0176, "step": 125450 }, { "epoch": 12.63, "learning_rate": 1.848759205084233e-05, "loss": 0.0136, "step": 125500 }, { "epoch": 12.63, "eval_cer": 0.062254434959121786, "eval_loss": 0.5363407135009766, "eval_runtime": 1624.4397, "eval_samples_per_second": 5.02, "eval_steps_per_second": 0.628, "step": 125500 }, { "epoch": 12.63, "learning_rate": 1.84749823464138e-05, "loss": 0.0173, "step": 125550 }, { "epoch": 12.64, "learning_rate": 1.8462372641985273e-05, "loss": 0.0154, "step": 125600 }, { "epoch": 12.64, "learning_rate": 1.8449762937556744e-05, "loss": 0.016, "step": 125650 }, { "epoch": 12.65, "learning_rate": 1.8437153233128216e-05, "loss": 0.0187, "step": 125700 }, { "epoch": 12.65, "learning_rate": 1.842454352869969e-05, "loss": 0.0147, "step": 125750 }, { "epoch": 12.66, "learning_rate": 1.841193382427116e-05, "loss": 0.0177, "step": 125800 }, { "epoch": 12.66, "learning_rate": 1.839932411984263e-05, "loss": 0.0153, "step": 125850 }, { "epoch": 12.67, "learning_rate": 1.83867144154141e-05, "loss": 0.0147, "step": 125900 }, { "epoch": 12.67, "learning_rate": 1.8374104710985577e-05, "loss": 0.0165, "step": 125950 }, { "epoch": 12.68, "learning_rate": 1.8361495006557048e-05, "loss": 0.0152, "step": 126000 }, { "epoch": 12.68, "eval_cer": 0.06294559741966015, "eval_loss": 0.5364506244659424, "eval_runtime": 1638.8882, "eval_samples_per_second": 4.976, "eval_steps_per_second": 0.622, "step": 126000 }, { "epoch": 12.68, "learning_rate": 1.834888530212852e-05, "loss": 0.0167, "step": 126050 }, { "epoch": 12.69, "learning_rate": 1.833627559769999e-05, "loss": 0.0185, "step": 126100 }, { "epoch": 12.69, "learning_rate": 1.8323665893271462e-05, "loss": 0.013, "step": 126150 }, { "epoch": 12.7, "learning_rate": 1.8311056188842934e-05, "loss": 0.0156, "step": 126200 }, { "epoch": 12.7, "learning_rate": 1.8298446484414405e-05, "loss": 0.0171, "step": 126250 }, { "epoch": 12.71, "learning_rate": 1.828583677998588e-05, "loss": 0.016, "step": 126300 }, { "epoch": 12.71, "learning_rate": 1.8273227075557352e-05, "loss": 0.0141, "step": 126350 }, { "epoch": 12.72, "learning_rate": 1.826061737112882e-05, "loss": 0.0149, "step": 126400 }, { "epoch": 12.72, "learning_rate": 1.824800766670029e-05, "loss": 0.0176, "step": 126450 }, { "epoch": 12.73, "learning_rate": 1.8235397962271766e-05, "loss": 0.0153, "step": 126500 }, { "epoch": 12.73, "eval_cer": 0.06202716081669251, "eval_loss": 0.5368106961250305, "eval_runtime": 1333.6205, "eval_samples_per_second": 6.115, "eval_steps_per_second": 0.765, "step": 126500 }, { "epoch": 12.73, "learning_rate": 1.8222788257843238e-05, "loss": 0.0174, "step": 126550 }, { "epoch": 12.74, "learning_rate": 1.821017855341471e-05, "loss": 0.0166, "step": 126600 }, { "epoch": 12.74, "learning_rate": 1.819782104307475e-05, "loss": 0.018, "step": 126650 }, { "epoch": 12.75, "learning_rate": 1.8185211338646225e-05, "loss": 0.0147, "step": 126700 }, { "epoch": 12.75, "learning_rate": 1.8172601634217697e-05, "loss": 0.0194, "step": 126750 }, { "epoch": 12.76, "learning_rate": 1.8159991929789165e-05, "loss": 0.0145, "step": 126800 }, { "epoch": 12.76, "learning_rate": 1.8147382225360636e-05, "loss": 0.0176, "step": 126850 }, { "epoch": 12.77, "learning_rate": 1.813477252093211e-05, "loss": 0.0174, "step": 126900 }, { "epoch": 12.77, "learning_rate": 1.8122162816503583e-05, "loss": 0.0169, "step": 126950 }, { "epoch": 12.78, "learning_rate": 1.8109553112075054e-05, "loss": 0.0192, "step": 127000 }, { "epoch": 12.78, "eval_cer": 0.06212678783803137, "eval_loss": 0.5365011692047119, "eval_runtime": 1312.139, "eval_samples_per_second": 6.215, "eval_steps_per_second": 0.777, "step": 127000 }, { "epoch": 12.78, "learning_rate": 1.8096943407646526e-05, "loss": 0.0149, "step": 127050 }, { "epoch": 12.79, "learning_rate": 1.8084333703217997e-05, "loss": 0.0168, "step": 127100 }, { "epoch": 12.79, "learning_rate": 1.807172399878947e-05, "loss": 0.017, "step": 127150 }, { "epoch": 12.8, "learning_rate": 1.805911429436094e-05, "loss": 0.0165, "step": 127200 }, { "epoch": 12.8, "learning_rate": 1.8046504589932415e-05, "loss": 0.0158, "step": 127250 }, { "epoch": 12.81, "learning_rate": 1.8033894885503887e-05, "loss": 0.0143, "step": 127300 }, { "epoch": 12.81, "learning_rate": 1.8021285181075355e-05, "loss": 0.019, "step": 127350 }, { "epoch": 12.82, "learning_rate": 1.8008675476646826e-05, "loss": 0.0148, "step": 127400 }, { "epoch": 12.82, "learning_rate": 1.79960657722183e-05, "loss": 0.0204, "step": 127450 }, { "epoch": 12.83, "learning_rate": 1.7983456067789773e-05, "loss": 0.0156, "step": 127500 }, { "epoch": 12.83, "eval_cer": 0.062341608602793294, "eval_loss": 0.5356007814407349, "eval_runtime": 1314.7354, "eval_samples_per_second": 6.203, "eval_steps_per_second": 0.776, "step": 127500 }, { "epoch": 12.83, "learning_rate": 1.7970846363361244e-05, "loss": 0.0137, "step": 127550 }, { "epoch": 12.84, "learning_rate": 1.7958236658932715e-05, "loss": 0.0147, "step": 127600 }, { "epoch": 12.84, "learning_rate": 1.7945626954504187e-05, "loss": 0.0188, "step": 127650 }, { "epoch": 12.85, "learning_rate": 1.793326944416423e-05, "loss": 0.0172, "step": 127700 }, { "epoch": 12.85, "learning_rate": 1.79206597397357e-05, "loss": 0.0171, "step": 127750 }, { "epoch": 12.86, "learning_rate": 1.790805003530717e-05, "loss": 0.015, "step": 127800 }, { "epoch": 12.86, "learning_rate": 1.7895440330878646e-05, "loss": 0.0169, "step": 127850 }, { "epoch": 12.87, "learning_rate": 1.7882830626450118e-05, "loss": 0.017, "step": 127900 }, { "epoch": 12.87, "learning_rate": 1.787022092202159e-05, "loss": 0.0162, "step": 127950 }, { "epoch": 12.88, "learning_rate": 1.785761121759306e-05, "loss": 0.0176, "step": 128000 }, { "epoch": 12.88, "eval_cer": 0.061989800683690435, "eval_loss": 0.538143515586853, "eval_runtime": 1313.0406, "eval_samples_per_second": 6.211, "eval_steps_per_second": 0.777, "step": 128000 }, { "epoch": 12.88, "learning_rate": 1.7845001513164532e-05, "loss": 0.0173, "step": 128050 }, { "epoch": 12.89, "learning_rate": 1.7832391808736004e-05, "loss": 0.0165, "step": 128100 }, { "epoch": 12.89, "learning_rate": 1.7819782104307475e-05, "loss": 0.0174, "step": 128150 }, { "epoch": 12.9, "learning_rate": 1.7807172399878946e-05, "loss": 0.0193, "step": 128200 }, { "epoch": 12.9, "learning_rate": 1.779456269545042e-05, "loss": 0.0182, "step": 128250 }, { "epoch": 12.91, "learning_rate": 1.7781952991021893e-05, "loss": 0.0163, "step": 128300 }, { "epoch": 12.92, "learning_rate": 1.776934328659336e-05, "loss": 0.0157, "step": 128350 }, { "epoch": 12.92, "learning_rate": 1.7756733582164836e-05, "loss": 0.0162, "step": 128400 }, { "epoch": 12.93, "learning_rate": 1.7744123877736307e-05, "loss": 0.0156, "step": 128450 }, { "epoch": 12.93, "learning_rate": 1.773151417330778e-05, "loss": 0.0172, "step": 128500 }, { "epoch": 12.93, "eval_cer": 0.06218282803753448, "eval_loss": 0.5344347357749939, "eval_runtime": 1650.819, "eval_samples_per_second": 4.94, "eval_steps_per_second": 0.618, "step": 128500 }, { "epoch": 12.94, "learning_rate": 1.771890446887925e-05, "loss": 0.0167, "step": 128550 }, { "epoch": 12.94, "learning_rate": 1.7706294764450722e-05, "loss": 0.0158, "step": 128600 }, { "epoch": 12.95, "learning_rate": 1.7693685060022193e-05, "loss": 0.0165, "step": 128650 }, { "epoch": 12.95, "learning_rate": 1.7681075355593665e-05, "loss": 0.017, "step": 128700 }, { "epoch": 12.96, "learning_rate": 1.7668465651165136e-05, "loss": 0.0173, "step": 128750 }, { "epoch": 12.96, "learning_rate": 1.765585594673661e-05, "loss": 0.022, "step": 128800 }, { "epoch": 12.97, "learning_rate": 1.7643246242308083e-05, "loss": 0.0139, "step": 128850 }, { "epoch": 12.97, "learning_rate": 1.7630636537879554e-05, "loss": 0.0149, "step": 128900 }, { "epoch": 12.98, "learning_rate": 1.7618026833451022e-05, "loss": 0.0174, "step": 128950 }, { "epoch": 12.98, "learning_rate": 1.7605417129022497e-05, "loss": 0.0134, "step": 129000 }, { "epoch": 12.98, "eval_cer": 0.06251284254571947, "eval_loss": 0.5331001281738281, "eval_runtime": 1681.0529, "eval_samples_per_second": 4.851, "eval_steps_per_second": 0.607, "step": 129000 }, { "epoch": 12.99, "learning_rate": 1.759280742459397e-05, "loss": 0.0165, "step": 129050 }, { "epoch": 12.99, "learning_rate": 1.758019772016544e-05, "loss": 0.0185, "step": 129100 }, { "epoch": 13.0, "learning_rate": 1.7567588015736915e-05, "loss": 0.0154, "step": 129150 }, { "epoch": 13.0, "learning_rate": 1.7554978311308383e-05, "loss": 0.0146, "step": 129200 }, { "epoch": 13.01, "learning_rate": 1.7542368606879854e-05, "loss": 0.0103, "step": 129250 }, { "epoch": 13.01, "learning_rate": 1.7529758902451326e-05, "loss": 0.011, "step": 129300 }, { "epoch": 13.02, "learning_rate": 1.75171491980228e-05, "loss": 0.0108, "step": 129350 }, { "epoch": 13.02, "learning_rate": 1.7504539493594272e-05, "loss": 0.0111, "step": 129400 }, { "epoch": 13.03, "learning_rate": 1.7491929789165744e-05, "loss": 0.0091, "step": 129450 }, { "epoch": 13.03, "learning_rate": 1.7479320084737212e-05, "loss": 0.0095, "step": 129500 }, { "epoch": 13.03, "eval_cer": 0.06145430544399405, "eval_loss": 0.5352779626846313, "eval_runtime": 1656.1399, "eval_samples_per_second": 4.924, "eval_steps_per_second": 0.616, "step": 129500 }, { "epoch": 13.04, "learning_rate": 1.7466710380308687e-05, "loss": 0.0115, "step": 129550 }, { "epoch": 13.04, "learning_rate": 1.7454100675880158e-05, "loss": 0.0107, "step": 129600 }, { "epoch": 13.05, "learning_rate": 1.744149097145163e-05, "loss": 0.0075, "step": 129650 }, { "epoch": 13.05, "learning_rate": 1.74288812670231e-05, "loss": 0.0085, "step": 129700 }, { "epoch": 13.06, "learning_rate": 1.7416271562594576e-05, "loss": 0.01, "step": 129750 }, { "epoch": 13.06, "learning_rate": 1.7403661858166044e-05, "loss": 0.0102, "step": 129800 }, { "epoch": 13.07, "learning_rate": 1.7391052153737516e-05, "loss": 0.009, "step": 129850 }, { "epoch": 13.07, "learning_rate": 1.737844244930899e-05, "loss": 0.0099, "step": 129900 }, { "epoch": 13.08, "learning_rate": 1.7365832744880462e-05, "loss": 0.0092, "step": 129950 }, { "epoch": 13.08, "learning_rate": 1.7353223040451933e-05, "loss": 0.0117, "step": 130000 }, { "epoch": 13.08, "eval_cer": 0.0619555538951052, "eval_loss": 0.5373625755310059, "eval_runtime": 1563.7692, "eval_samples_per_second": 5.215, "eval_steps_per_second": 0.652, "step": 130000 }, { "epoch": 13.09, "learning_rate": 1.7340613336023405e-05, "loss": 0.0115, "step": 130050 }, { "epoch": 13.09, "learning_rate": 1.7328003631594876e-05, "loss": 0.0096, "step": 130100 }, { "epoch": 13.1, "learning_rate": 1.7315393927166348e-05, "loss": 0.0118, "step": 130150 }, { "epoch": 13.1, "learning_rate": 1.730278422273782e-05, "loss": 0.0126, "step": 130200 }, { "epoch": 13.11, "learning_rate": 1.729017451830929e-05, "loss": 0.0101, "step": 130250 }, { "epoch": 13.11, "learning_rate": 1.7277564813880766e-05, "loss": 0.0124, "step": 130300 }, { "epoch": 13.12, "learning_rate": 1.7264955109452234e-05, "loss": 0.0088, "step": 130350 }, { "epoch": 13.12, "learning_rate": 1.7252345405023705e-05, "loss": 0.0101, "step": 130400 }, { "epoch": 13.13, "learning_rate": 1.7239735700595177e-05, "loss": 0.0084, "step": 130450 }, { "epoch": 13.13, "learning_rate": 1.722712599616665e-05, "loss": 0.0102, "step": 130500 }, { "epoch": 13.13, "eval_cer": 0.06132665832290363, "eval_loss": 0.5378304123878479, "eval_runtime": 1599.7805, "eval_samples_per_second": 5.098, "eval_steps_per_second": 0.638, "step": 130500 }, { "epoch": 13.14, "learning_rate": 1.7214516291738123e-05, "loss": 0.0086, "step": 130550 }, { "epoch": 13.14, "learning_rate": 1.7201906587309595e-05, "loss": 0.0132, "step": 130600 }, { "epoch": 13.15, "learning_rate": 1.7189296882881066e-05, "loss": 0.0163, "step": 130650 }, { "epoch": 13.15, "learning_rate": 1.7176687178452538e-05, "loss": 0.0094, "step": 130700 }, { "epoch": 13.16, "learning_rate": 1.716407747402401e-05, "loss": 0.0111, "step": 130750 }, { "epoch": 13.16, "learning_rate": 1.715146776959548e-05, "loss": 0.0104, "step": 130800 }, { "epoch": 13.17, "learning_rate": 1.7138858065166955e-05, "loss": 0.0106, "step": 130850 }, { "epoch": 13.17, "learning_rate": 1.7126248360738427e-05, "loss": 0.0085, "step": 130900 }, { "epoch": 13.18, "learning_rate": 1.7113638656309895e-05, "loss": 0.0133, "step": 130950 }, { "epoch": 13.18, "learning_rate": 1.7101028951881367e-05, "loss": 0.009, "step": 131000 }, { "epoch": 13.18, "eval_cer": 0.06143251203307617, "eval_loss": 0.5404409766197205, "eval_runtime": 1315.2853, "eval_samples_per_second": 6.2, "eval_steps_per_second": 0.775, "step": 131000 }, { "epoch": 13.19, "learning_rate": 1.708841924745284e-05, "loss": 0.0099, "step": 131050 }, { "epoch": 13.19, "learning_rate": 1.7075809543024313e-05, "loss": 0.0098, "step": 131100 }, { "epoch": 13.2, "learning_rate": 1.7063199838595784e-05, "loss": 0.01, "step": 131150 }, { "epoch": 13.2, "learning_rate": 1.7050590134167256e-05, "loss": 0.0105, "step": 131200 }, { "epoch": 13.21, "learning_rate": 1.7037980429738727e-05, "loss": 0.0122, "step": 131250 }, { "epoch": 13.21, "learning_rate": 1.70253707253102e-05, "loss": 0.0104, "step": 131300 }, { "epoch": 13.22, "learning_rate": 1.701276102088167e-05, "loss": 0.008, "step": 131350 }, { "epoch": 13.22, "learning_rate": 1.7000151316453145e-05, "loss": 0.0095, "step": 131400 }, { "epoch": 13.23, "learning_rate": 1.6987541612024617e-05, "loss": 0.0102, "step": 131450 }, { "epoch": 13.23, "learning_rate": 1.6974931907596088e-05, "loss": 0.0101, "step": 131500 }, { "epoch": 13.23, "eval_cer": 0.06098107709263445, "eval_loss": 0.5369338393211365, "eval_runtime": 1307.2678, "eval_samples_per_second": 6.238, "eval_steps_per_second": 0.78, "step": 131500 }, { "epoch": 13.24, "learning_rate": 1.6962322203167556e-05, "loss": 0.0095, "step": 131550 }, { "epoch": 13.24, "learning_rate": 1.694971249873903e-05, "loss": 0.0122, "step": 131600 }, { "epoch": 13.25, "learning_rate": 1.6937102794310503e-05, "loss": 0.0132, "step": 131650 }, { "epoch": 13.25, "learning_rate": 1.6924493089881974e-05, "loss": 0.0092, "step": 131700 }, { "epoch": 13.26, "learning_rate": 1.6911883385453446e-05, "loss": 0.0099, "step": 131750 }, { "epoch": 13.26, "learning_rate": 1.6899273681024917e-05, "loss": 0.0101, "step": 131800 }, { "epoch": 13.27, "learning_rate": 1.688666397659639e-05, "loss": 0.0118, "step": 131850 }, { "epoch": 13.27, "learning_rate": 1.687405427216786e-05, "loss": 0.0114, "step": 131900 }, { "epoch": 13.28, "learning_rate": 1.686144456773933e-05, "loss": 0.0124, "step": 131950 }, { "epoch": 13.28, "learning_rate": 1.6848834863310806e-05, "loss": 0.01, "step": 132000 }, { "epoch": 13.28, "eval_cer": 0.06100909719238601, "eval_loss": 0.5384798049926758, "eval_runtime": 1315.6412, "eval_samples_per_second": 6.198, "eval_steps_per_second": 0.775, "step": 132000 }, { "epoch": 13.29, "learning_rate": 1.6836225158882278e-05, "loss": 0.0094, "step": 132050 }, { "epoch": 13.29, "learning_rate": 1.682361545445375e-05, "loss": 0.0099, "step": 132100 }, { "epoch": 13.3, "learning_rate": 1.681100575002522e-05, "loss": 0.0101, "step": 132150 }, { "epoch": 13.3, "learning_rate": 1.6798396045596692e-05, "loss": 0.0106, "step": 132200 }, { "epoch": 13.31, "learning_rate": 1.6785786341168164e-05, "loss": 0.0096, "step": 132250 }, { "epoch": 13.31, "learning_rate": 1.6773176636739635e-05, "loss": 0.0111, "step": 132300 }, { "epoch": 13.32, "learning_rate": 1.676056693231111e-05, "loss": 0.0086, "step": 132350 }, { "epoch": 13.32, "learning_rate": 1.6747957227882578e-05, "loss": 0.0083, "step": 132400 }, { "epoch": 13.33, "learning_rate": 1.673534752345405e-05, "loss": 0.0123, "step": 132450 }, { "epoch": 13.33, "learning_rate": 1.672273781902552e-05, "loss": 0.0112, "step": 132500 }, { "epoch": 13.33, "eval_cer": 0.06100909719238601, "eval_loss": 0.5370435118675232, "eval_runtime": 1366.383, "eval_samples_per_second": 5.968, "eval_steps_per_second": 0.746, "step": 132500 }, { "epoch": 13.34, "learning_rate": 1.6710128114596996e-05, "loss": 0.0129, "step": 132550 }, { "epoch": 13.34, "learning_rate": 1.6697518410168468e-05, "loss": 0.0134, "step": 132600 }, { "epoch": 13.35, "learning_rate": 1.668490870573994e-05, "loss": 0.0104, "step": 132650 }, { "epoch": 13.35, "learning_rate": 1.667229900131141e-05, "loss": 0.012, "step": 132700 }, { "epoch": 13.36, "learning_rate": 1.6659689296882882e-05, "loss": 0.0105, "step": 132750 }, { "epoch": 13.36, "learning_rate": 1.6647079592454353e-05, "loss": 0.0105, "step": 132800 }, { "epoch": 13.37, "learning_rate": 1.6634469888025825e-05, "loss": 0.0103, "step": 132850 }, { "epoch": 13.37, "learning_rate": 1.6621860183597296e-05, "loss": 0.0089, "step": 132900 }, { "epoch": 13.38, "learning_rate": 1.660925047916877e-05, "loss": 0.0092, "step": 132950 }, { "epoch": 13.38, "learning_rate": 1.659664077474024e-05, "loss": 0.009, "step": 133000 }, { "epoch": 13.38, "eval_cer": 0.06120212454623005, "eval_loss": 0.5355344414710999, "eval_runtime": 1322.6487, "eval_samples_per_second": 6.166, "eval_steps_per_second": 0.771, "step": 133000 }, { "epoch": 13.39, "learning_rate": 1.658403107031171e-05, "loss": 0.0105, "step": 133050 }, { "epoch": 13.39, "learning_rate": 1.6571421365883186e-05, "loss": 0.0094, "step": 133100 }, { "epoch": 13.4, "learning_rate": 1.6559063855543227e-05, "loss": 0.0103, "step": 133150 }, { "epoch": 13.4, "learning_rate": 1.65464541511147e-05, "loss": 0.011, "step": 133200 }, { "epoch": 13.41, "learning_rate": 1.653384444668617e-05, "loss": 0.0115, "step": 133250 }, { "epoch": 13.41, "learning_rate": 1.6521234742257645e-05, "loss": 0.0106, "step": 133300 }, { "epoch": 13.42, "learning_rate": 1.6508625037829113e-05, "loss": 0.0103, "step": 133350 }, { "epoch": 13.42, "learning_rate": 1.6496015333400584e-05, "loss": 0.0113, "step": 133400 }, { "epoch": 13.43, "learning_rate": 1.6483405628972056e-05, "loss": 0.0108, "step": 133450 }, { "epoch": 13.43, "learning_rate": 1.647079592454353e-05, "loss": 0.0108, "step": 133500 }, { "epoch": 13.43, "eval_cer": 0.06121457792389741, "eval_loss": 0.541231632232666, "eval_runtime": 1318.2465, "eval_samples_per_second": 6.186, "eval_steps_per_second": 0.774, "step": 133500 }, { "epoch": 13.44, "learning_rate": 1.6458186220115002e-05, "loss": 0.0093, "step": 133550 }, { "epoch": 13.44, "learning_rate": 1.6445576515686474e-05, "loss": 0.0115, "step": 133600 }, { "epoch": 13.45, "learning_rate": 1.6432966811257945e-05, "loss": 0.0125, "step": 133650 }, { "epoch": 13.45, "learning_rate": 1.6420357106829417e-05, "loss": 0.0141, "step": 133700 }, { "epoch": 13.46, "learning_rate": 1.6407747402400888e-05, "loss": 0.0111, "step": 133750 }, { "epoch": 13.46, "learning_rate": 1.639513769797236e-05, "loss": 0.0091, "step": 133800 }, { "epoch": 13.47, "learning_rate": 1.638252799354383e-05, "loss": 0.0099, "step": 133850 }, { "epoch": 13.47, "learning_rate": 1.6369918289115306e-05, "loss": 0.0129, "step": 133900 }, { "epoch": 13.48, "learning_rate": 1.6357308584686774e-05, "loss": 0.0119, "step": 133950 }, { "epoch": 13.48, "learning_rate": 1.6344698880258246e-05, "loss": 0.0091, "step": 134000 }, { "epoch": 13.48, "eval_cer": 0.06103089060330388, "eval_loss": 0.5410909056663513, "eval_runtime": 1351.917, "eval_samples_per_second": 6.032, "eval_steps_per_second": 0.754, "step": 134000 }, { "epoch": 13.49, "learning_rate": 1.633208917582972e-05, "loss": 0.0112, "step": 134050 }, { "epoch": 13.49, "learning_rate": 1.6319479471401192e-05, "loss": 0.0095, "step": 134100 }, { "epoch": 13.5, "learning_rate": 1.6306869766972664e-05, "loss": 0.0124, "step": 134150 }, { "epoch": 13.5, "learning_rate": 1.6294260062544135e-05, "loss": 0.0109, "step": 134200 }, { "epoch": 13.51, "learning_rate": 1.6281650358115606e-05, "loss": 0.0097, "step": 134250 }, { "epoch": 13.51, "learning_rate": 1.6269040653687078e-05, "loss": 0.0138, "step": 134300 }, { "epoch": 13.52, "learning_rate": 1.625643094925855e-05, "loss": 0.0097, "step": 134350 }, { "epoch": 13.52, "learning_rate": 1.624382124483002e-05, "loss": 0.0087, "step": 134400 }, { "epoch": 13.53, "learning_rate": 1.6231211540401496e-05, "loss": 0.0109, "step": 134450 }, { "epoch": 13.53, "learning_rate": 1.6218601835972967e-05, "loss": 0.0111, "step": 134500 }, { "epoch": 13.53, "eval_cer": 0.06137335848915622, "eval_loss": 0.5402405261993408, "eval_runtime": 1314.6644, "eval_samples_per_second": 6.203, "eval_steps_per_second": 0.776, "step": 134500 }, { "epoch": 13.54, "learning_rate": 1.6205992131544435e-05, "loss": 0.0112, "step": 134550 }, { "epoch": 13.54, "learning_rate": 1.6193382427115907e-05, "loss": 0.0133, "step": 134600 }, { "epoch": 13.55, "learning_rate": 1.6180772722687382e-05, "loss": 0.0109, "step": 134650 }, { "epoch": 13.55, "learning_rate": 1.6168163018258853e-05, "loss": 0.0098, "step": 134700 }, { "epoch": 13.56, "learning_rate": 1.6155553313830325e-05, "loss": 0.0141, "step": 134750 }, { "epoch": 13.56, "learning_rate": 1.6142943609401796e-05, "loss": 0.0131, "step": 134800 }, { "epoch": 13.57, "learning_rate": 1.6130333904973268e-05, "loss": 0.0106, "step": 134850 }, { "epoch": 13.57, "learning_rate": 1.611772420054474e-05, "loss": 0.0102, "step": 134900 }, { "epoch": 13.58, "learning_rate": 1.610511449611621e-05, "loss": 0.0093, "step": 134950 }, { "epoch": 13.58, "learning_rate": 1.6092504791687686e-05, "loss": 0.01, "step": 135000 }, { "epoch": 13.58, "eval_cer": 0.06106513739188912, "eval_loss": 0.5380209684371948, "eval_runtime": 1313.4736, "eval_samples_per_second": 6.209, "eval_steps_per_second": 0.777, "step": 135000 }, { "epoch": 13.59, "learning_rate": 1.6079895087259157e-05, "loss": 0.0131, "step": 135050 }, { "epoch": 13.59, "learning_rate": 1.606728538283063e-05, "loss": 0.0111, "step": 135100 }, { "epoch": 13.6, "learning_rate": 1.6054675678402097e-05, "loss": 0.0113, "step": 135150 }, { "epoch": 13.6, "learning_rate": 1.604206597397357e-05, "loss": 0.012, "step": 135200 }, { "epoch": 13.61, "learning_rate": 1.6029456269545043e-05, "loss": 0.0099, "step": 135250 }, { "epoch": 13.61, "learning_rate": 1.6016846565116514e-05, "loss": 0.0103, "step": 135300 }, { "epoch": 13.62, "learning_rate": 1.6004236860687986e-05, "loss": 0.012, "step": 135350 }, { "epoch": 13.62, "learning_rate": 1.5991627156259457e-05, "loss": 0.0094, "step": 135400 }, { "epoch": 13.63, "learning_rate": 1.597901745183093e-05, "loss": 0.0093, "step": 135450 }, { "epoch": 13.63, "learning_rate": 1.59664077474024e-05, "loss": 0.0109, "step": 135500 }, { "epoch": 13.63, "eval_cer": 0.061547705776499234, "eval_loss": 0.5380915403366089, "eval_runtime": 1315.2144, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 135500 }, { "epoch": 13.64, "learning_rate": 1.5953798042973875e-05, "loss": 0.0107, "step": 135550 }, { "epoch": 13.64, "learning_rate": 1.5941188338545347e-05, "loss": 0.0122, "step": 135600 }, { "epoch": 13.65, "learning_rate": 1.5928578634116818e-05, "loss": 0.0114, "step": 135650 }, { "epoch": 13.65, "learning_rate": 1.591596892968829e-05, "loss": 0.0115, "step": 135700 }, { "epoch": 13.66, "learning_rate": 1.590361141934833e-05, "loss": 0.0115, "step": 135750 }, { "epoch": 13.66, "learning_rate": 1.5891001714919802e-05, "loss": 0.012, "step": 135800 }, { "epoch": 13.67, "learning_rate": 1.5878392010491274e-05, "loss": 0.0116, "step": 135850 }, { "epoch": 13.67, "learning_rate": 1.5865782306062745e-05, "loss": 0.0098, "step": 135900 }, { "epoch": 13.68, "learning_rate": 1.585317260163422e-05, "loss": 0.0124, "step": 135950 }, { "epoch": 13.68, "learning_rate": 1.5840562897205692e-05, "loss": 0.0102, "step": 136000 }, { "epoch": 13.68, "eval_cer": 0.06052652880777589, "eval_loss": 0.5393700003623962, "eval_runtime": 1315.1677, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 136000 }, { "epoch": 13.69, "learning_rate": 1.5827953192777163e-05, "loss": 0.0087, "step": 136050 }, { "epoch": 13.69, "learning_rate": 1.581534348834863e-05, "loss": 0.0081, "step": 136100 }, { "epoch": 13.7, "learning_rate": 1.5802733783920106e-05, "loss": 0.0107, "step": 136150 }, { "epoch": 13.7, "learning_rate": 1.5790124079491578e-05, "loss": 0.0104, "step": 136200 }, { "epoch": 13.71, "learning_rate": 1.577751437506305e-05, "loss": 0.0111, "step": 136250 }, { "epoch": 13.71, "learning_rate": 1.576490467063452e-05, "loss": 0.0115, "step": 136300 }, { "epoch": 13.72, "learning_rate": 1.5752294966205992e-05, "loss": 0.0118, "step": 136350 }, { "epoch": 13.73, "learning_rate": 1.5739685261777464e-05, "loss": 0.0108, "step": 136400 }, { "epoch": 13.73, "learning_rate": 1.5727075557348935e-05, "loss": 0.0099, "step": 136450 }, { "epoch": 13.74, "learning_rate": 1.5714465852920407e-05, "loss": 0.0108, "step": 136500 }, { "epoch": 13.74, "eval_cer": 0.06086588334921139, "eval_loss": 0.5379374623298645, "eval_runtime": 1323.8185, "eval_samples_per_second": 6.16, "eval_steps_per_second": 0.77, "step": 136500 }, { "epoch": 13.74, "learning_rate": 1.570185614849188e-05, "loss": 0.0153, "step": 136550 }, { "epoch": 13.75, "learning_rate": 1.5689246444063353e-05, "loss": 0.0094, "step": 136600 }, { "epoch": 13.75, "learning_rate": 1.5676636739634824e-05, "loss": 0.009, "step": 136650 }, { "epoch": 13.76, "learning_rate": 1.5664027035206296e-05, "loss": 0.0098, "step": 136700 }, { "epoch": 13.76, "learning_rate": 1.5651417330777767e-05, "loss": 0.0093, "step": 136750 }, { "epoch": 13.77, "learning_rate": 1.563880762634924e-05, "loss": 0.0102, "step": 136800 }, { "epoch": 13.77, "learning_rate": 1.562619792192071e-05, "loss": 0.0098, "step": 136850 }, { "epoch": 13.78, "learning_rate": 1.5613588217492185e-05, "loss": 0.0088, "step": 136900 }, { "epoch": 13.78, "learning_rate": 1.5600978513063653e-05, "loss": 0.0099, "step": 136950 }, { "epoch": 13.79, "learning_rate": 1.5588368808635125e-05, "loss": 0.0088, "step": 137000 }, { "epoch": 13.79, "eval_cer": 0.06057634231844532, "eval_loss": 0.5392413139343262, "eval_runtime": 1335.0393, "eval_samples_per_second": 6.108, "eval_steps_per_second": 0.764, "step": 137000 }, { "epoch": 13.79, "learning_rate": 1.5575759104206596e-05, "loss": 0.0102, "step": 137050 }, { "epoch": 13.8, "learning_rate": 1.556314939977807e-05, "loss": 0.0114, "step": 137100 }, { "epoch": 13.8, "learning_rate": 1.5550539695349543e-05, "loss": 0.0091, "step": 137150 }, { "epoch": 13.81, "learning_rate": 1.5537929990921014e-05, "loss": 0.0226, "step": 137200 }, { "epoch": 13.81, "learning_rate": 1.5525320286492486e-05, "loss": 0.0125, "step": 137250 }, { "epoch": 13.82, "learning_rate": 1.5512710582063957e-05, "loss": 0.0105, "step": 137300 }, { "epoch": 13.82, "learning_rate": 1.550010087763543e-05, "loss": 0.0095, "step": 137350 }, { "epoch": 13.83, "learning_rate": 1.54874911732069e-05, "loss": 0.0118, "step": 137400 }, { "epoch": 13.83, "learning_rate": 1.5474881468778375e-05, "loss": 0.0115, "step": 137450 }, { "epoch": 13.84, "learning_rate": 1.5462271764349846e-05, "loss": 0.011, "step": 137500 }, { "epoch": 13.84, "eval_cer": 0.06133288501173731, "eval_loss": 0.5379477739334106, "eval_runtime": 1315.0915, "eval_samples_per_second": 6.201, "eval_steps_per_second": 0.776, "step": 137500 }, { "epoch": 13.84, "learning_rate": 1.5449662059921315e-05, "loss": 0.014, "step": 137550 }, { "epoch": 13.85, "learning_rate": 1.5437052355492786e-05, "loss": 0.0104, "step": 137600 }, { "epoch": 13.85, "learning_rate": 1.542444265106426e-05, "loss": 0.01, "step": 137650 }, { "epoch": 13.86, "learning_rate": 1.5411832946635732e-05, "loss": 0.0116, "step": 137700 }, { "epoch": 13.86, "learning_rate": 1.5399223242207204e-05, "loss": 0.0103, "step": 137750 }, { "epoch": 13.87, "learning_rate": 1.5386613537778675e-05, "loss": 0.0122, "step": 137800 }, { "epoch": 13.87, "learning_rate": 1.5374003833350147e-05, "loss": 0.0102, "step": 137850 }, { "epoch": 13.88, "learning_rate": 1.536139412892162e-05, "loss": 0.0094, "step": 137900 }, { "epoch": 13.88, "learning_rate": 1.534878442449309e-05, "loss": 0.0113, "step": 137950 }, { "epoch": 13.89, "learning_rate": 1.533617472006456e-05, "loss": 0.0152, "step": 138000 }, { "epoch": 13.89, "eval_cer": 0.060766256327872525, "eval_loss": 0.5388569831848145, "eval_runtime": 1336.8781, "eval_samples_per_second": 6.1, "eval_steps_per_second": 0.763, "step": 138000 }, { "epoch": 13.89, "learning_rate": 1.5323565015636036e-05, "loss": 0.0132, "step": 138050 }, { "epoch": 13.9, "learning_rate": 1.5310955311207508e-05, "loss": 0.0121, "step": 138100 }, { "epoch": 13.9, "learning_rate": 1.5298345606778976e-05, "loss": 0.011, "step": 138150 }, { "epoch": 13.91, "learning_rate": 1.528573590235045e-05, "loss": 0.0128, "step": 138200 }, { "epoch": 13.91, "learning_rate": 1.5273126197921922e-05, "loss": 0.0108, "step": 138250 }, { "epoch": 13.92, "learning_rate": 1.5260516493493394e-05, "loss": 0.0083, "step": 138300 }, { "epoch": 13.92, "learning_rate": 1.5247906789064863e-05, "loss": 0.0125, "step": 138350 }, { "epoch": 13.93, "learning_rate": 1.5235297084636338e-05, "loss": 0.0107, "step": 138400 }, { "epoch": 13.93, "learning_rate": 1.522268738020781e-05, "loss": 0.0099, "step": 138450 }, { "epoch": 13.94, "learning_rate": 1.521007767577928e-05, "loss": 0.0113, "step": 138500 }, { "epoch": 13.94, "eval_cer": 0.06074757626137149, "eval_loss": 0.5370819568634033, "eval_runtime": 1308.527, "eval_samples_per_second": 6.232, "eval_steps_per_second": 0.78, "step": 138500 }, { "epoch": 13.94, "learning_rate": 1.5197467971350751e-05, "loss": 0.0109, "step": 138550 }, { "epoch": 13.95, "learning_rate": 1.5184858266922226e-05, "loss": 0.0127, "step": 138600 }, { "epoch": 13.95, "learning_rate": 1.5172500756582267e-05, "loss": 0.0136, "step": 138650 }, { "epoch": 13.96, "learning_rate": 1.5159891052153739e-05, "loss": 0.0108, "step": 138700 }, { "epoch": 13.96, "learning_rate": 1.5147281347725208e-05, "loss": 0.011, "step": 138750 }, { "epoch": 13.97, "learning_rate": 1.5134671643296683e-05, "loss": 0.0135, "step": 138800 }, { "epoch": 13.97, "learning_rate": 1.5122061938868153e-05, "loss": 0.0092, "step": 138850 }, { "epoch": 13.98, "learning_rate": 1.5109452234439625e-05, "loss": 0.0113, "step": 138900 }, { "epoch": 13.98, "learning_rate": 1.5096842530011096e-05, "loss": 0.0124, "step": 138950 }, { "epoch": 13.99, "learning_rate": 1.508423282558257e-05, "loss": 0.0105, "step": 139000 }, { "epoch": 13.99, "eval_cer": 0.06119901120181321, "eval_loss": 0.5397846698760986, "eval_runtime": 1308.762, "eval_samples_per_second": 6.231, "eval_steps_per_second": 0.779, "step": 139000 }, { "epoch": 13.99, "learning_rate": 1.507162312115404e-05, "loss": 0.0098, "step": 139050 }, { "epoch": 14.0, "learning_rate": 1.5059013416725512e-05, "loss": 0.0112, "step": 139100 }, { "epoch": 14.0, "learning_rate": 1.5046403712296985e-05, "loss": 0.0122, "step": 139150 }, { "epoch": 14.01, "learning_rate": 1.5033794007868457e-05, "loss": 0.007, "step": 139200 }, { "epoch": 14.01, "learning_rate": 1.5021184303439928e-05, "loss": 0.0075, "step": 139250 }, { "epoch": 14.02, "learning_rate": 1.50085745990114e-05, "loss": 0.0061, "step": 139300 }, { "epoch": 14.02, "learning_rate": 1.4995964894582873e-05, "loss": 0.0068, "step": 139350 }, { "epoch": 14.03, "learning_rate": 1.4983355190154345e-05, "loss": 0.0087, "step": 139400 }, { "epoch": 14.03, "learning_rate": 1.4970745485725814e-05, "loss": 0.0072, "step": 139450 }, { "epoch": 14.04, "learning_rate": 1.4958135781297286e-05, "loss": 0.0094, "step": 139500 }, { "epoch": 14.04, "eval_cer": 0.060327274765098166, "eval_loss": 0.5419248938560486, "eval_runtime": 1313.2281, "eval_samples_per_second": 6.21, "eval_steps_per_second": 0.777, "step": 139500 }, { "epoch": 14.04, "learning_rate": 1.494552607686876e-05, "loss": 0.0071, "step": 139550 }, { "epoch": 14.05, "learning_rate": 1.493291637244023e-05, "loss": 0.0073, "step": 139600 }, { "epoch": 14.05, "learning_rate": 1.4920306668011702e-05, "loss": 0.0058, "step": 139650 }, { "epoch": 14.06, "learning_rate": 1.4907696963583173e-05, "loss": 0.0073, "step": 139700 }, { "epoch": 14.06, "learning_rate": 1.4895087259154647e-05, "loss": 0.0082, "step": 139750 }, { "epoch": 14.07, "learning_rate": 1.4882477554726118e-05, "loss": 0.0079, "step": 139800 }, { "epoch": 14.07, "learning_rate": 1.486986785029759e-05, "loss": 0.0047, "step": 139850 }, { "epoch": 14.08, "learning_rate": 1.485725814586906e-05, "loss": 0.007, "step": 139900 }, { "epoch": 14.08, "learning_rate": 1.4844648441440534e-05, "loss": 0.0062, "step": 139950 }, { "epoch": 14.09, "learning_rate": 1.4832038737012006e-05, "loss": 0.0073, "step": 140000 }, { "epoch": 14.09, "eval_cer": 0.06108381745839015, "eval_loss": 0.543082594871521, "eval_runtime": 1322.7308, "eval_samples_per_second": 6.165, "eval_steps_per_second": 0.771, "step": 140000 }, { "epoch": 14.09, "learning_rate": 1.4819429032583475e-05, "loss": 0.0071, "step": 140050 }, { "epoch": 14.1, "learning_rate": 1.480681932815495e-05, "loss": 0.0081, "step": 140100 }, { "epoch": 14.1, "learning_rate": 1.4794209623726422e-05, "loss": 0.0083, "step": 140150 }, { "epoch": 14.11, "learning_rate": 1.4781599919297892e-05, "loss": 0.007, "step": 140200 }, { "epoch": 14.11, "learning_rate": 1.4768990214869363e-05, "loss": 0.0062, "step": 140250 }, { "epoch": 14.12, "learning_rate": 1.4756380510440836e-05, "loss": 0.007, "step": 140300 }, { "epoch": 14.12, "learning_rate": 1.4743770806012308e-05, "loss": 0.0077, "step": 140350 }, { "epoch": 14.13, "learning_rate": 1.473116110158378e-05, "loss": 0.006, "step": 140400 }, { "epoch": 14.13, "learning_rate": 1.471855139715525e-05, "loss": 0.0063, "step": 140450 }, { "epoch": 14.14, "learning_rate": 1.4705941692726724e-05, "loss": 0.0094, "step": 140500 }, { "epoch": 14.14, "eval_cer": 0.06001282697899738, "eval_loss": 0.5420997738838196, "eval_runtime": 1510.2178, "eval_samples_per_second": 5.4, "eval_steps_per_second": 0.675, "step": 140500 }, { "epoch": 14.14, "learning_rate": 1.4693331988298195e-05, "loss": 0.0065, "step": 140550 }, { "epoch": 14.15, "learning_rate": 1.4680722283869667e-05, "loss": 0.0073, "step": 140600 }, { "epoch": 14.15, "learning_rate": 1.4668112579441137e-05, "loss": 0.0088, "step": 140650 }, { "epoch": 14.16, "learning_rate": 1.4655502875012612e-05, "loss": 0.0066, "step": 140700 }, { "epoch": 14.16, "learning_rate": 1.4642893170584081e-05, "loss": 0.0077, "step": 140750 }, { "epoch": 14.17, "learning_rate": 1.4630283466155553e-05, "loss": 0.0052, "step": 140800 }, { "epoch": 14.17, "learning_rate": 1.4617673761727028e-05, "loss": 0.0064, "step": 140850 }, { "epoch": 14.18, "learning_rate": 1.4605064057298497e-05, "loss": 0.0056, "step": 140900 }, { "epoch": 14.18, "learning_rate": 1.4592454352869969e-05, "loss": 0.0076, "step": 140950 }, { "epoch": 14.19, "learning_rate": 1.457984464844144e-05, "loss": 0.0058, "step": 141000 }, { "epoch": 14.19, "eval_cer": 0.06005952714524997, "eval_loss": 0.5442749857902527, "eval_runtime": 1624.0101, "eval_samples_per_second": 5.022, "eval_steps_per_second": 0.628, "step": 141000 }, { "epoch": 14.19, "learning_rate": 1.4567234944012914e-05, "loss": 0.0063, "step": 141050 }, { "epoch": 14.2, "learning_rate": 1.4554625239584385e-05, "loss": 0.0065, "step": 141100 }, { "epoch": 14.2, "learning_rate": 1.4542015535155857e-05, "loss": 0.0056, "step": 141150 }, { "epoch": 14.21, "learning_rate": 1.4529405830727328e-05, "loss": 0.0068, "step": 141200 }, { "epoch": 14.21, "learning_rate": 1.4516796126298801e-05, "loss": 0.0073, "step": 141250 }, { "epoch": 14.22, "learning_rate": 1.4504186421870273e-05, "loss": 0.0075, "step": 141300 }, { "epoch": 14.22, "learning_rate": 1.4491576717441743e-05, "loss": 0.0052, "step": 141350 }, { "epoch": 14.23, "learning_rate": 1.4478967013013214e-05, "loss": 0.0085, "step": 141400 }, { "epoch": 14.23, "learning_rate": 1.4466357308584689e-05, "loss": 0.008, "step": 141450 }, { "epoch": 14.24, "learning_rate": 1.4453747604156159e-05, "loss": 0.0076, "step": 141500 }, { "epoch": 14.24, "eval_cer": 0.0597855528365681, "eval_loss": 0.5461323261260986, "eval_runtime": 1434.8268, "eval_samples_per_second": 5.684, "eval_steps_per_second": 0.711, "step": 141500 }, { "epoch": 14.24, "learning_rate": 1.444113789972763e-05, "loss": 0.0055, "step": 141550 }, { "epoch": 14.25, "learning_rate": 1.4428528195299103e-05, "loss": 0.0075, "step": 141600 }, { "epoch": 14.25, "learning_rate": 1.4415918490870575e-05, "loss": 0.0066, "step": 141650 }, { "epoch": 14.26, "learning_rate": 1.4403308786442046e-05, "loss": 0.0059, "step": 141700 }, { "epoch": 14.26, "learning_rate": 1.4390951276102088e-05, "loss": 0.007, "step": 141750 }, { "epoch": 14.27, "learning_rate": 1.4378341571673562e-05, "loss": 0.006, "step": 141800 }, { "epoch": 14.27, "learning_rate": 1.4365731867245032e-05, "loss": 0.0074, "step": 141850 }, { "epoch": 14.28, "learning_rate": 1.4353122162816504e-05, "loss": 0.0063, "step": 141900 }, { "epoch": 14.28, "learning_rate": 1.4340512458387975e-05, "loss": 0.0085, "step": 141950 }, { "epoch": 14.29, "learning_rate": 1.4327902753959448e-05, "loss": 0.0082, "step": 142000 }, { "epoch": 14.29, "eval_cer": 0.05971705925939763, "eval_loss": 0.5430257320404053, "eval_runtime": 1312.6088, "eval_samples_per_second": 6.213, "eval_steps_per_second": 0.777, "step": 142000 }, { "epoch": 14.29, "learning_rate": 1.431529304953092e-05, "loss": 0.0065, "step": 142050 }, { "epoch": 14.3, "learning_rate": 1.4302683345102391e-05, "loss": 0.0084, "step": 142100 }, { "epoch": 14.3, "learning_rate": 1.4290073640673863e-05, "loss": 0.0059, "step": 142150 }, { "epoch": 14.31, "learning_rate": 1.4277463936245336e-05, "loss": 0.0054, "step": 142200 }, { "epoch": 14.31, "learning_rate": 1.4264854231816808e-05, "loss": 0.0066, "step": 142250 }, { "epoch": 14.32, "learning_rate": 1.4252244527388279e-05, "loss": 0.007, "step": 142300 }, { "epoch": 14.32, "learning_rate": 1.4239634822959749e-05, "loss": 0.0063, "step": 142350 }, { "epoch": 14.33, "learning_rate": 1.4227025118531224e-05, "loss": 0.0063, "step": 142400 }, { "epoch": 14.33, "learning_rate": 1.4214415414102693e-05, "loss": 0.0069, "step": 142450 }, { "epoch": 14.34, "learning_rate": 1.4201805709674165e-05, "loss": 0.0067, "step": 142500 }, { "epoch": 14.34, "eval_cer": 0.060840976593876675, "eval_loss": 0.5442531704902649, "eval_runtime": 1311.1335, "eval_samples_per_second": 6.22, "eval_steps_per_second": 0.778, "step": 142500 }, { "epoch": 14.34, "learning_rate": 1.418919600524564e-05, "loss": 0.0064, "step": 142550 }, { "epoch": 14.35, "learning_rate": 1.417658630081711e-05, "loss": 0.0067, "step": 142600 }, { "epoch": 14.35, "learning_rate": 1.4163976596388581e-05, "loss": 0.0104, "step": 142650 }, { "epoch": 14.36, "learning_rate": 1.4151366891960053e-05, "loss": 0.0056, "step": 142700 }, { "epoch": 14.36, "learning_rate": 1.4138757187531526e-05, "loss": 0.0071, "step": 142750 }, { "epoch": 14.37, "learning_rate": 1.4126147483102997e-05, "loss": 0.0088, "step": 142800 }, { "epoch": 14.37, "learning_rate": 1.4113537778674469e-05, "loss": 0.0075, "step": 142850 }, { "epoch": 14.38, "learning_rate": 1.4100928074245938e-05, "loss": 0.0089, "step": 142900 }, { "epoch": 14.38, "learning_rate": 1.4088318369817413e-05, "loss": 0.0086, "step": 142950 }, { "epoch": 14.39, "learning_rate": 1.4075708665388885e-05, "loss": 0.0055, "step": 143000 }, { "epoch": 14.39, "eval_cer": 0.05994122005741007, "eval_loss": 0.5452024936676025, "eval_runtime": 1318.3262, "eval_samples_per_second": 6.186, "eval_steps_per_second": 0.774, "step": 143000 }, { "epoch": 14.39, "learning_rate": 1.4063098960960355e-05, "loss": 0.0076, "step": 143050 }, { "epoch": 14.4, "learning_rate": 1.4050489256531826e-05, "loss": 0.0054, "step": 143100 }, { "epoch": 14.4, "learning_rate": 1.4037879552103301e-05, "loss": 0.006, "step": 143150 }, { "epoch": 14.41, "learning_rate": 1.402526984767477e-05, "loss": 0.0069, "step": 143200 }, { "epoch": 14.41, "learning_rate": 1.4012660143246242e-05, "loss": 0.0075, "step": 143250 }, { "epoch": 14.42, "learning_rate": 1.4000050438817715e-05, "loss": 0.0064, "step": 143300 }, { "epoch": 14.42, "learning_rate": 1.3987440734389187e-05, "loss": 0.0063, "step": 143350 }, { "epoch": 14.43, "learning_rate": 1.3974831029960658e-05, "loss": 0.0067, "step": 143400 }, { "epoch": 14.43, "learning_rate": 1.396222132553213e-05, "loss": 0.0066, "step": 143450 }, { "epoch": 14.44, "learning_rate": 1.3949611621103603e-05, "loss": 0.0068, "step": 143500 }, { "epoch": 14.44, "eval_cer": 0.060327274765098166, "eval_loss": 0.5434070825576782, "eval_runtime": 1642.6331, "eval_samples_per_second": 4.965, "eval_steps_per_second": 0.621, "step": 143500 }, { "epoch": 14.44, "learning_rate": 1.3937001916675075e-05, "loss": 0.0057, "step": 143550 }, { "epoch": 14.45, "learning_rate": 1.3924392212246546e-05, "loss": 0.009, "step": 143600 }, { "epoch": 14.45, "learning_rate": 1.3911782507818016e-05, "loss": 0.0071, "step": 143650 }, { "epoch": 14.46, "learning_rate": 1.389917280338949e-05, "loss": 0.009, "step": 143700 }, { "epoch": 14.46, "learning_rate": 1.388656309896096e-05, "loss": 0.0071, "step": 143750 }, { "epoch": 14.47, "learning_rate": 1.3873953394532432e-05, "loss": 0.0066, "step": 143800 }, { "epoch": 14.47, "learning_rate": 1.3861343690103903e-05, "loss": 0.0072, "step": 143850 }, { "epoch": 14.48, "learning_rate": 1.3848733985675377e-05, "loss": 0.0076, "step": 143900 }, { "epoch": 14.48, "learning_rate": 1.3836124281246848e-05, "loss": 0.0068, "step": 143950 }, { "epoch": 14.49, "learning_rate": 1.382351457681832e-05, "loss": 0.0066, "step": 144000 }, { "epoch": 14.49, "eval_cer": 0.060268121221178216, "eval_loss": 0.5480110049247742, "eval_runtime": 1649.6672, "eval_samples_per_second": 4.943, "eval_steps_per_second": 0.618, "step": 144000 }, { "epoch": 14.49, "learning_rate": 1.3810904872389791e-05, "loss": 0.0065, "step": 144050 }, { "epoch": 14.5, "learning_rate": 1.3798295167961264e-05, "loss": 0.0066, "step": 144100 }, { "epoch": 14.5, "learning_rate": 1.3785685463532736e-05, "loss": 0.0068, "step": 144150 }, { "epoch": 14.51, "learning_rate": 1.3773075759104207e-05, "loss": 0.0053, "step": 144200 }, { "epoch": 14.51, "learning_rate": 1.376046605467568e-05, "loss": 0.0067, "step": 144250 }, { "epoch": 14.52, "learning_rate": 1.3747856350247152e-05, "loss": 0.0067, "step": 144300 }, { "epoch": 14.53, "learning_rate": 1.3735246645818622e-05, "loss": 0.0084, "step": 144350 }, { "epoch": 14.53, "learning_rate": 1.3722636941390093e-05, "loss": 0.0071, "step": 144400 }, { "epoch": 14.54, "learning_rate": 1.3710027236961568e-05, "loss": 0.0084, "step": 144450 }, { "epoch": 14.54, "learning_rate": 1.3697417532533038e-05, "loss": 0.0064, "step": 144500 }, { "epoch": 14.54, "eval_cer": 0.06016849419983935, "eval_loss": 0.5481834411621094, "eval_runtime": 1635.9379, "eval_samples_per_second": 4.985, "eval_steps_per_second": 0.623, "step": 144500 }, { "epoch": 14.55, "learning_rate": 1.368480782810451e-05, "loss": 0.0068, "step": 144550 }, { "epoch": 14.55, "learning_rate": 1.367219812367598e-05, "loss": 0.0058, "step": 144600 }, { "epoch": 14.56, "learning_rate": 1.3659588419247454e-05, "loss": 0.0076, "step": 144650 }, { "epoch": 14.56, "learning_rate": 1.3646978714818925e-05, "loss": 0.0072, "step": 144700 }, { "epoch": 14.57, "learning_rate": 1.3634369010390397e-05, "loss": 0.0076, "step": 144750 }, { "epoch": 14.57, "learning_rate": 1.3621759305961867e-05, "loss": 0.0076, "step": 144800 }, { "epoch": 14.58, "learning_rate": 1.3609149601533342e-05, "loss": 0.0065, "step": 144850 }, { "epoch": 14.58, "learning_rate": 1.3596792091193383e-05, "loss": 0.0077, "step": 144900 }, { "epoch": 14.59, "learning_rate": 1.3584182386764854e-05, "loss": 0.0082, "step": 144950 }, { "epoch": 14.59, "learning_rate": 1.3571572682336326e-05, "loss": 0.0073, "step": 145000 }, { "epoch": 14.59, "eval_cer": 0.0601249073780036, "eval_loss": 0.5465731024742126, "eval_runtime": 1480.8252, "eval_samples_per_second": 5.507, "eval_steps_per_second": 0.689, "step": 145000 }, { "epoch": 14.6, "learning_rate": 1.3558962977907799e-05, "loss": 0.0071, "step": 145050 }, { "epoch": 14.6, "learning_rate": 1.354635327347927e-05, "loss": 0.0066, "step": 145100 }, { "epoch": 14.61, "learning_rate": 1.3533743569050742e-05, "loss": 0.0064, "step": 145150 }, { "epoch": 14.61, "learning_rate": 1.3521133864622215e-05, "loss": 0.01, "step": 145200 }, { "epoch": 14.62, "learning_rate": 1.3508524160193687e-05, "loss": 0.0072, "step": 145250 }, { "epoch": 14.62, "learning_rate": 1.3495914455765158e-05, "loss": 0.0071, "step": 145300 }, { "epoch": 14.63, "learning_rate": 1.3483304751336628e-05, "loss": 0.0076, "step": 145350 }, { "epoch": 14.63, "learning_rate": 1.3470695046908103e-05, "loss": 0.0069, "step": 145400 }, { "epoch": 14.64, "learning_rate": 1.3458085342479573e-05, "loss": 0.0068, "step": 145450 }, { "epoch": 14.64, "learning_rate": 1.3445475638051044e-05, "loss": 0.0081, "step": 145500 }, { "epoch": 14.64, "eval_cer": 0.06027434791001189, "eval_loss": 0.5453492403030396, "eval_runtime": 1628.7583, "eval_samples_per_second": 5.007, "eval_steps_per_second": 0.626, "step": 145500 }, { "epoch": 14.65, "learning_rate": 1.3432865933622516e-05, "loss": 0.0062, "step": 145550 }, { "epoch": 14.65, "learning_rate": 1.3420256229193989e-05, "loss": 0.0074, "step": 145600 }, { "epoch": 14.66, "learning_rate": 1.340764652476546e-05, "loss": 0.0065, "step": 145650 }, { "epoch": 14.66, "learning_rate": 1.3395036820336932e-05, "loss": 0.0064, "step": 145700 }, { "epoch": 14.67, "learning_rate": 1.3382427115908403e-05, "loss": 0.0052, "step": 145750 }, { "epoch": 14.67, "learning_rate": 1.3369817411479876e-05, "loss": 0.0059, "step": 145800 }, { "epoch": 14.68, "learning_rate": 1.3357207707051348e-05, "loss": 0.0068, "step": 145850 }, { "epoch": 14.68, "learning_rate": 1.3344598002622818e-05, "loss": 0.0065, "step": 145900 }, { "epoch": 14.69, "learning_rate": 1.3331988298194293e-05, "loss": 0.0061, "step": 145950 }, { "epoch": 14.69, "learning_rate": 1.3319378593765764e-05, "loss": 0.007, "step": 146000 }, { "epoch": 14.69, "eval_cer": 0.060280574598845574, "eval_loss": 0.5450297594070435, "eval_runtime": 1318.9659, "eval_samples_per_second": 6.183, "eval_steps_per_second": 0.773, "step": 146000 }, { "epoch": 14.7, "learning_rate": 1.3306768889337234e-05, "loss": 0.0069, "step": 146050 }, { "epoch": 14.7, "learning_rate": 1.3294159184908705e-05, "loss": 0.0073, "step": 146100 }, { "epoch": 14.71, "learning_rate": 1.328154948048018e-05, "loss": 0.0098, "step": 146150 }, { "epoch": 14.71, "learning_rate": 1.326893977605165e-05, "loss": 0.0067, "step": 146200 }, { "epoch": 14.72, "learning_rate": 1.3256330071623121e-05, "loss": 0.0063, "step": 146250 }, { "epoch": 14.72, "learning_rate": 1.3243720367194593e-05, "loss": 0.006, "step": 146300 }, { "epoch": 14.73, "learning_rate": 1.3231110662766066e-05, "loss": 0.0084, "step": 146350 }, { "epoch": 14.73, "learning_rate": 1.3218500958337538e-05, "loss": 0.0063, "step": 146400 }, { "epoch": 14.74, "learning_rate": 1.3205891253909009e-05, "loss": 0.0073, "step": 146450 }, { "epoch": 14.74, "learning_rate": 1.3193281549480479e-05, "loss": 0.0066, "step": 146500 }, { "epoch": 14.74, "eval_cer": 0.06055454890752744, "eval_loss": 0.5488942265510559, "eval_runtime": 1316.8458, "eval_samples_per_second": 6.193, "eval_steps_per_second": 0.775, "step": 146500 }, { "epoch": 14.75, "learning_rate": 1.3180671845051954e-05, "loss": 0.0057, "step": 146550 }, { "epoch": 14.75, "learning_rate": 1.3168062140623425e-05, "loss": 0.0086, "step": 146600 }, { "epoch": 14.76, "learning_rate": 1.3155452436194895e-05, "loss": 0.0061, "step": 146650 }, { "epoch": 14.76, "learning_rate": 1.314284273176637e-05, "loss": 0.0067, "step": 146700 }, { "epoch": 14.77, "learning_rate": 1.313023302733784e-05, "loss": 0.0072, "step": 146750 }, { "epoch": 14.77, "learning_rate": 1.3117875516997883e-05, "loss": 0.0094, "step": 146800 }, { "epoch": 14.78, "learning_rate": 1.3105265812569354e-05, "loss": 0.0068, "step": 146850 }, { "epoch": 14.78, "learning_rate": 1.3092656108140824e-05, "loss": 0.0065, "step": 146900 }, { "epoch": 14.79, "learning_rate": 1.3080046403712299e-05, "loss": 0.0054, "step": 146950 }, { "epoch": 14.79, "learning_rate": 1.3067436699283769e-05, "loss": 0.0056, "step": 147000 }, { "epoch": 14.79, "eval_cer": 0.05963922564897664, "eval_loss": 0.5449287295341492, "eval_runtime": 1314.5286, "eval_samples_per_second": 6.204, "eval_steps_per_second": 0.776, "step": 147000 }, { "epoch": 14.8, "learning_rate": 1.305482699485524e-05, "loss": 0.0084, "step": 147050 }, { "epoch": 14.8, "learning_rate": 1.3042217290426715e-05, "loss": 0.0076, "step": 147100 }, { "epoch": 14.81, "learning_rate": 1.3029607585998185e-05, "loss": 0.0078, "step": 147150 }, { "epoch": 14.81, "learning_rate": 1.3016997881569656e-05, "loss": 0.0068, "step": 147200 }, { "epoch": 14.82, "learning_rate": 1.3004388177141128e-05, "loss": 0.0047, "step": 147250 }, { "epoch": 14.82, "learning_rate": 1.2991778472712601e-05, "loss": 0.0063, "step": 147300 }, { "epoch": 14.83, "learning_rate": 1.2979168768284072e-05, "loss": 0.0089, "step": 147350 }, { "epoch": 14.83, "learning_rate": 1.2966559063855544e-05, "loss": 0.0064, "step": 147400 }, { "epoch": 14.84, "learning_rate": 1.2953949359427014e-05, "loss": 0.0191, "step": 147450 }, { "epoch": 14.84, "learning_rate": 1.2941339654998489e-05, "loss": 0.0076, "step": 147500 }, { "epoch": 14.84, "eval_cer": 0.05977932614773442, "eval_loss": 0.5438098907470703, "eval_runtime": 1317.9364, "eval_samples_per_second": 6.188, "eval_steps_per_second": 0.774, "step": 147500 }, { "epoch": 14.85, "learning_rate": 1.292872995056996e-05, "loss": 0.0072, "step": 147550 }, { "epoch": 14.85, "learning_rate": 1.291612024614143e-05, "loss": 0.0074, "step": 147600 }, { "epoch": 14.86, "learning_rate": 1.2903510541712901e-05, "loss": 0.0062, "step": 147650 }, { "epoch": 14.86, "learning_rate": 1.2890900837284376e-05, "loss": 0.0074, "step": 147700 }, { "epoch": 14.87, "learning_rate": 1.2878291132855846e-05, "loss": 0.006, "step": 147750 }, { "epoch": 14.87, "learning_rate": 1.2865681428427317e-05, "loss": 0.0068, "step": 147800 }, { "epoch": 14.88, "learning_rate": 1.285307172399879e-05, "loss": 0.0062, "step": 147850 }, { "epoch": 14.88, "learning_rate": 1.2840462019570262e-05, "loss": 0.0064, "step": 147900 }, { "epoch": 14.89, "learning_rate": 1.2827852315141734e-05, "loss": 0.0066, "step": 147950 }, { "epoch": 14.89, "learning_rate": 1.2815242610713205e-05, "loss": 0.0092, "step": 148000 }, { "epoch": 14.89, "eval_cer": 0.059147317231116006, "eval_loss": 0.5454152226448059, "eval_runtime": 1312.4041, "eval_samples_per_second": 6.214, "eval_steps_per_second": 0.777, "step": 148000 }, { "epoch": 14.9, "learning_rate": 1.2802632906284678e-05, "loss": 0.0062, "step": 148050 }, { "epoch": 14.9, "learning_rate": 1.279002320185615e-05, "loss": 0.0077, "step": 148100 }, { "epoch": 14.91, "learning_rate": 1.2777413497427621e-05, "loss": 0.0066, "step": 148150 }, { "epoch": 14.91, "learning_rate": 1.2764803792999091e-05, "loss": 0.0061, "step": 148200 }, { "epoch": 14.92, "learning_rate": 1.2752194088570566e-05, "loss": 0.0092, "step": 148250 }, { "epoch": 14.92, "learning_rate": 1.2739584384142036e-05, "loss": 0.0071, "step": 148300 }, { "epoch": 14.93, "learning_rate": 1.2726974679713507e-05, "loss": 0.0066, "step": 148350 }, { "epoch": 14.93, "learning_rate": 1.2714364975284979e-05, "loss": 0.007, "step": 148400 }, { "epoch": 14.94, "learning_rate": 1.2701755270856452e-05, "loss": 0.0066, "step": 148450 }, { "epoch": 14.94, "learning_rate": 1.2689145566427923e-05, "loss": 0.01, "step": 148500 }, { "epoch": 14.94, "eval_cer": 0.059754419392399706, "eval_loss": 0.543897807598114, "eval_runtime": 1311.7963, "eval_samples_per_second": 6.217, "eval_steps_per_second": 0.778, "step": 148500 }, { "epoch": 14.95, "learning_rate": 1.2676535861999395e-05, "loss": 0.006, "step": 148550 }, { "epoch": 14.95, "learning_rate": 1.2663926157570868e-05, "loss": 0.006, "step": 148600 }, { "epoch": 14.96, "learning_rate": 1.265131645314234e-05, "loss": 0.0078, "step": 148650 }, { "epoch": 14.96, "learning_rate": 1.2638706748713811e-05, "loss": 0.0085, "step": 148700 }, { "epoch": 14.97, "learning_rate": 1.2626097044285282e-05, "loss": 0.0072, "step": 148750 }, { "epoch": 14.97, "learning_rate": 1.2613487339856756e-05, "loss": 0.0051, "step": 148800 }, { "epoch": 14.98, "learning_rate": 1.2600877635428227e-05, "loss": 0.0052, "step": 148850 }, { "epoch": 14.98, "learning_rate": 1.2588267930999697e-05, "loss": 0.0055, "step": 148900 }, { "epoch": 14.99, "learning_rate": 1.2575658226571168e-05, "loss": 0.0069, "step": 148950 }, { "epoch": 14.99, "learning_rate": 1.2563048522142643e-05, "loss": 0.0053, "step": 149000 }, { "epoch": 14.99, "eval_cer": 0.05912863716461497, "eval_loss": 0.5436471104621887, "eval_runtime": 1314.2835, "eval_samples_per_second": 6.205, "eval_steps_per_second": 0.776, "step": 149000 }, { "epoch": 15.0, "learning_rate": 1.2550438817714113e-05, "loss": 0.0092, "step": 149050 }, { "epoch": 15.0, "learning_rate": 1.2537829113285584e-05, "loss": 0.0057, "step": 149100 }, { "epoch": 15.01, "learning_rate": 1.2525219408857056e-05, "loss": 0.0043, "step": 149150 }, { "epoch": 15.01, "learning_rate": 1.2512609704428529e-05, "loss": 0.0036, "step": 149200 }, { "epoch": 15.02, "learning_rate": 1.25e-05, "loss": 0.0037, "step": 149250 }, { "epoch": 15.02, "learning_rate": 1.2487390295571474e-05, "loss": 0.0041, "step": 149300 }, { "epoch": 15.03, "learning_rate": 1.2474780591142944e-05, "loss": 0.0042, "step": 149350 }, { "epoch": 15.03, "learning_rate": 1.2462170886714417e-05, "loss": 0.004, "step": 149400 }, { "epoch": 15.04, "learning_rate": 1.2449561182285888e-05, "loss": 0.0034, "step": 149450 }, { "epoch": 15.04, "learning_rate": 1.243695147785736e-05, "loss": 0.0033, "step": 149500 }, { "epoch": 15.04, "eval_cer": 0.05933411789612638, "eval_loss": 0.5475030541419983, "eval_runtime": 1308.9478, "eval_samples_per_second": 6.23, "eval_steps_per_second": 0.779, "step": 149500 }, { "epoch": 15.05, "learning_rate": 1.2424341773428831e-05, "loss": 0.0053, "step": 149550 }, { "epoch": 15.05, "learning_rate": 1.2411732069000304e-05, "loss": 0.0046, "step": 149600 }, { "epoch": 15.06, "learning_rate": 1.2399122364571774e-05, "loss": 0.0043, "step": 149650 }, { "epoch": 15.06, "learning_rate": 1.2386512660143247e-05, "loss": 0.0031, "step": 149700 }, { "epoch": 15.07, "learning_rate": 1.2373902955714719e-05, "loss": 0.0042, "step": 149750 }, { "epoch": 15.07, "learning_rate": 1.236129325128619e-05, "loss": 0.0052, "step": 149800 }, { "epoch": 15.08, "learning_rate": 1.2348683546857662e-05, "loss": 0.0055, "step": 149850 }, { "epoch": 15.08, "learning_rate": 1.2336073842429135e-05, "loss": 0.0047, "step": 149900 }, { "epoch": 15.09, "learning_rate": 1.2323464138000605e-05, "loss": 0.0053, "step": 149950 }, { "epoch": 15.09, "learning_rate": 1.2310854433572078e-05, "loss": 0.005, "step": 150000 }, { "epoch": 15.09, "eval_cer": 0.05908193699836238, "eval_loss": 0.5471116900444031, "eval_runtime": 1311.7989, "eval_samples_per_second": 6.217, "eval_steps_per_second": 0.778, "step": 150000 }, { "epoch": 15.1, "learning_rate": 1.229824472914355e-05, "loss": 0.0062, "step": 150050 }, { "epoch": 15.1, "learning_rate": 1.2285635024715021e-05, "loss": 0.005, "step": 150100 }, { "epoch": 15.11, "learning_rate": 1.2273025320286494e-05, "loss": 0.004, "step": 150150 }, { "epoch": 15.11, "learning_rate": 1.2260415615857966e-05, "loss": 0.0048, "step": 150200 }, { "epoch": 15.12, "learning_rate": 1.2247805911429437e-05, "loss": 0.0035, "step": 150250 }, { "epoch": 15.12, "learning_rate": 1.2235196207000909e-05, "loss": 0.0048, "step": 150300 }, { "epoch": 15.13, "learning_rate": 1.222258650257238e-05, "loss": 0.0041, "step": 150350 }, { "epoch": 15.13, "learning_rate": 1.2209976798143851e-05, "loss": 0.0049, "step": 150400 }, { "epoch": 15.14, "learning_rate": 1.2197367093715325e-05, "loss": 0.0042, "step": 150450 }, { "epoch": 15.14, "learning_rate": 1.2184757389286794e-05, "loss": 0.0064, "step": 150500 }, { "epoch": 15.14, "eval_cer": 0.059754419392399706, "eval_loss": 0.5467462539672852, "eval_runtime": 1315.8444, "eval_samples_per_second": 6.198, "eval_steps_per_second": 0.775, "step": 150500 }, { "epoch": 15.15, "learning_rate": 1.2172147684858268e-05, "loss": 0.0052, "step": 150550 }, { "epoch": 15.15, "learning_rate": 1.2159537980429739e-05, "loss": 0.0053, "step": 150600 }, { "epoch": 15.16, "learning_rate": 1.214692827600121e-05, "loss": 0.0041, "step": 150650 }, { "epoch": 15.16, "learning_rate": 1.2134318571572682e-05, "loss": 0.0041, "step": 150700 }, { "epoch": 15.17, "learning_rate": 1.2121708867144155e-05, "loss": 0.0041, "step": 150750 }, { "epoch": 15.17, "learning_rate": 1.2109099162715625e-05, "loss": 0.0052, "step": 150800 }, { "epoch": 15.18, "learning_rate": 1.2096489458287098e-05, "loss": 0.0051, "step": 150850 }, { "epoch": 15.18, "learning_rate": 1.2083879753858571e-05, "loss": 0.0052, "step": 150900 }, { "epoch": 15.19, "learning_rate": 1.2071270049430041e-05, "loss": 0.004, "step": 150950 }, { "epoch": 15.19, "learning_rate": 1.2058660345001514e-05, "loss": 0.0046, "step": 151000 }, { "epoch": 15.19, "eval_cer": 0.05920024408620228, "eval_loss": 0.5482798218727112, "eval_runtime": 1311.8482, "eval_samples_per_second": 6.216, "eval_steps_per_second": 0.778, "step": 151000 }, { "epoch": 15.2, "learning_rate": 1.2046050640572986e-05, "loss": 0.0049, "step": 151050 }, { "epoch": 15.2, "learning_rate": 1.2033440936144457e-05, "loss": 0.0046, "step": 151100 }, { "epoch": 15.21, "learning_rate": 1.2020831231715929e-05, "loss": 0.0034, "step": 151150 }, { "epoch": 15.21, "learning_rate": 1.2008221527287402e-05, "loss": 0.0043, "step": 151200 }, { "epoch": 15.22, "learning_rate": 1.1995611822858872e-05, "loss": 0.0038, "step": 151250 }, { "epoch": 15.22, "learning_rate": 1.1983002118430345e-05, "loss": 0.0049, "step": 151300 }, { "epoch": 15.23, "learning_rate": 1.1970392414001816e-05, "loss": 0.0039, "step": 151350 }, { "epoch": 15.23, "learning_rate": 1.1957782709573288e-05, "loss": 0.0033, "step": 151400 }, { "epoch": 15.24, "learning_rate": 1.194517300514476e-05, "loss": 0.0036, "step": 151450 }, { "epoch": 15.24, "learning_rate": 1.1932563300716233e-05, "loss": 0.0032, "step": 151500 }, { "epoch": 15.24, "eval_cer": 0.0593434579293769, "eval_loss": 0.5473946928977966, "eval_runtime": 1307.3758, "eval_samples_per_second": 6.238, "eval_steps_per_second": 0.78, "step": 151500 }, { "epoch": 15.25, "learning_rate": 1.1919953596287702e-05, "loss": 0.0048, "step": 151550 }, { "epoch": 15.25, "learning_rate": 1.1907343891859176e-05, "loss": 0.004, "step": 151600 }, { "epoch": 15.26, "learning_rate": 1.1894734187430647e-05, "loss": 0.0051, "step": 151650 }, { "epoch": 15.26, "learning_rate": 1.188237667709069e-05, "loss": 0.0044, "step": 151700 }, { "epoch": 15.27, "learning_rate": 1.1869766972662162e-05, "loss": 0.0076, "step": 151750 }, { "epoch": 15.27, "learning_rate": 1.1857157268233633e-05, "loss": 0.0067, "step": 151800 }, { "epoch": 15.28, "learning_rate": 1.1844547563805106e-05, "loss": 0.0058, "step": 151850 }, { "epoch": 15.28, "learning_rate": 1.1831937859376576e-05, "loss": 0.0036, "step": 151900 }, { "epoch": 15.29, "learning_rate": 1.181932815494805e-05, "loss": 0.0064, "step": 151950 }, { "epoch": 15.29, "learning_rate": 1.180671845051952e-05, "loss": 0.0039, "step": 152000 }, { "epoch": 15.29, "eval_cer": 0.05925628428570539, "eval_loss": 0.5505883097648621, "eval_runtime": 1311.9962, "eval_samples_per_second": 6.216, "eval_steps_per_second": 0.777, "step": 152000 }, { "epoch": 15.3, "learning_rate": 1.1794108746090992e-05, "loss": 0.0049, "step": 152050 }, { "epoch": 15.3, "learning_rate": 1.1781499041662464e-05, "loss": 0.0058, "step": 152100 }, { "epoch": 15.31, "learning_rate": 1.1768889337233937e-05, "loss": 0.0048, "step": 152150 }, { "epoch": 15.31, "learning_rate": 1.1756279632805407e-05, "loss": 0.0121, "step": 152200 }, { "epoch": 15.32, "learning_rate": 1.174366992837688e-05, "loss": 0.0045, "step": 152250 }, { "epoch": 15.32, "learning_rate": 1.1731060223948351e-05, "loss": 0.0035, "step": 152300 }, { "epoch": 15.33, "learning_rate": 1.1718450519519823e-05, "loss": 0.0043, "step": 152350 }, { "epoch": 15.34, "learning_rate": 1.1705840815091294e-05, "loss": 0.0063, "step": 152400 }, { "epoch": 15.34, "learning_rate": 1.1693231110662767e-05, "loss": 0.0044, "step": 152450 }, { "epoch": 15.35, "learning_rate": 1.1680621406234237e-05, "loss": 0.004, "step": 152500 }, { "epoch": 15.35, "eval_cer": 0.05977309945890074, "eval_loss": 0.550814151763916, "eval_runtime": 1313.7936, "eval_samples_per_second": 6.207, "eval_steps_per_second": 0.776, "step": 152500 }, { "epoch": 15.35, "learning_rate": 1.166801170180571e-05, "loss": 0.0049, "step": 152550 }, { "epoch": 15.36, "learning_rate": 1.1655401997377182e-05, "loss": 0.0038, "step": 152600 }, { "epoch": 15.36, "learning_rate": 1.1642792292948653e-05, "loss": 0.0042, "step": 152650 }, { "epoch": 15.37, "learning_rate": 1.1630182588520127e-05, "loss": 0.0037, "step": 152700 }, { "epoch": 15.37, "learning_rate": 1.1617572884091598e-05, "loss": 0.005, "step": 152750 }, { "epoch": 15.38, "learning_rate": 1.160496317966307e-05, "loss": 0.0037, "step": 152800 }, { "epoch": 15.38, "learning_rate": 1.1592353475234541e-05, "loss": 0.0052, "step": 152850 }, { "epoch": 15.39, "learning_rate": 1.1579743770806014e-05, "loss": 0.005, "step": 152900 }, { "epoch": 15.39, "learning_rate": 1.1567134066377484e-05, "loss": 0.005, "step": 152950 }, { "epoch": 15.4, "learning_rate": 1.1554524361948957e-05, "loss": 0.0034, "step": 153000 }, { "epoch": 15.4, "eval_cer": 0.05976064608123338, "eval_loss": 0.5509442090988159, "eval_runtime": 1315.2981, "eval_samples_per_second": 6.2, "eval_steps_per_second": 0.775, "step": 153000 }, { "epoch": 15.4, "learning_rate": 1.1541914657520429e-05, "loss": 0.0039, "step": 153050 }, { "epoch": 15.41, "learning_rate": 1.15293049530919e-05, "loss": 0.0053, "step": 153100 }, { "epoch": 15.41, "learning_rate": 1.1516695248663372e-05, "loss": 0.004, "step": 153150 }, { "epoch": 15.42, "learning_rate": 1.1504085544234843e-05, "loss": 0.0048, "step": 153200 }, { "epoch": 15.42, "learning_rate": 1.1491475839806315e-05, "loss": 0.0039, "step": 153250 }, { "epoch": 15.43, "learning_rate": 1.1478866135377788e-05, "loss": 0.0037, "step": 153300 }, { "epoch": 15.43, "learning_rate": 1.146625643094926e-05, "loss": 0.0061, "step": 153350 }, { "epoch": 15.44, "learning_rate": 1.145364672652073e-05, "loss": 0.0046, "step": 153400 }, { "epoch": 15.44, "learning_rate": 1.1441037022092204e-05, "loss": 0.0049, "step": 153450 }, { "epoch": 15.45, "learning_rate": 1.1428427317663674e-05, "loss": 0.0033, "step": 153500 }, { "epoch": 15.45, "eval_cer": 0.05961431889364193, "eval_loss": 0.5518068075180054, "eval_runtime": 1314.4035, "eval_samples_per_second": 6.204, "eval_steps_per_second": 0.776, "step": 153500 }, { "epoch": 15.45, "learning_rate": 1.1415817613235147e-05, "loss": 0.0043, "step": 153550 }, { "epoch": 15.46, "learning_rate": 1.1403207908806618e-05, "loss": 0.0048, "step": 153600 }, { "epoch": 15.46, "learning_rate": 1.139059820437809e-05, "loss": 0.0056, "step": 153650 }, { "epoch": 15.47, "learning_rate": 1.1377988499949561e-05, "loss": 0.004, "step": 153700 }, { "epoch": 15.47, "learning_rate": 1.1365378795521034e-05, "loss": 0.0044, "step": 153750 }, { "epoch": 15.48, "learning_rate": 1.1352769091092504e-05, "loss": 0.0041, "step": 153800 }, { "epoch": 15.48, "learning_rate": 1.1340159386663977e-05, "loss": 0.0038, "step": 153850 }, { "epoch": 15.49, "learning_rate": 1.1327549682235449e-05, "loss": 0.0052, "step": 153900 }, { "epoch": 15.49, "learning_rate": 1.131493997780692e-05, "loss": 0.0055, "step": 153950 }, { "epoch": 15.5, "learning_rate": 1.1302330273378392e-05, "loss": 0.0037, "step": 154000 }, { "epoch": 15.5, "eval_cer": 0.0595645053829725, "eval_loss": 0.5507568120956421, "eval_runtime": 1308.4724, "eval_samples_per_second": 6.232, "eval_steps_per_second": 0.78, "step": 154000 }, { "epoch": 15.5, "learning_rate": 1.1289720568949865e-05, "loss": 0.0044, "step": 154050 }, { "epoch": 15.51, "learning_rate": 1.1277110864521335e-05, "loss": 0.0035, "step": 154100 }, { "epoch": 15.51, "learning_rate": 1.1264501160092808e-05, "loss": 0.0038, "step": 154150 }, { "epoch": 15.52, "learning_rate": 1.1251891455664281e-05, "loss": 0.0038, "step": 154200 }, { "epoch": 15.52, "learning_rate": 1.1239281751235751e-05, "loss": 0.0048, "step": 154250 }, { "epoch": 15.53, "learning_rate": 1.1226672046807224e-05, "loss": 0.0036, "step": 154300 }, { "epoch": 15.53, "learning_rate": 1.1214062342378696e-05, "loss": 0.0042, "step": 154350 }, { "epoch": 15.54, "learning_rate": 1.1201452637950167e-05, "loss": 0.0041, "step": 154400 }, { "epoch": 15.54, "learning_rate": 1.1188842933521639e-05, "loss": 0.0034, "step": 154450 }, { "epoch": 15.55, "learning_rate": 1.1176233229093112e-05, "loss": 0.0036, "step": 154500 }, { "epoch": 15.55, "eval_cer": 0.05964233899339348, "eval_loss": 0.5511408448219299, "eval_runtime": 1310.4121, "eval_samples_per_second": 6.223, "eval_steps_per_second": 0.778, "step": 154500 }, { "epoch": 15.55, "learning_rate": 1.1163623524664582e-05, "loss": 0.0049, "step": 154550 }, { "epoch": 15.56, "learning_rate": 1.1151013820236055e-05, "loss": 0.0043, "step": 154600 }, { "epoch": 15.56, "learning_rate": 1.1138404115807526e-05, "loss": 0.0037, "step": 154650 }, { "epoch": 15.57, "learning_rate": 1.1125794411378998e-05, "loss": 0.003, "step": 154700 }, { "epoch": 15.57, "learning_rate": 1.111318470695047e-05, "loss": 0.0039, "step": 154750 }, { "epoch": 15.58, "learning_rate": 1.1100575002521942e-05, "loss": 0.0031, "step": 154800 }, { "epoch": 15.58, "learning_rate": 1.1087965298093412e-05, "loss": 0.003, "step": 154850 }, { "epoch": 15.59, "learning_rate": 1.1075355593664885e-05, "loss": 0.0046, "step": 154900 }, { "epoch": 15.59, "learning_rate": 1.1062745889236357e-05, "loss": 0.0037, "step": 154950 }, { "epoch": 15.6, "learning_rate": 1.1050136184807828e-05, "loss": 0.0035, "step": 155000 }, { "epoch": 15.6, "eval_cer": 0.05918156401970125, "eval_loss": 0.5500566363334656, "eval_runtime": 1310.4808, "eval_samples_per_second": 6.223, "eval_steps_per_second": 0.778, "step": 155000 }, { "epoch": 15.6, "learning_rate": 1.1037526480379301e-05, "loss": 0.0039, "step": 155050 }, { "epoch": 15.61, "learning_rate": 1.1024916775950771e-05, "loss": 0.004, "step": 155100 }, { "epoch": 15.61, "learning_rate": 1.1012307071522244e-05, "loss": 0.0042, "step": 155150 }, { "epoch": 15.62, "learning_rate": 1.0999697367093716e-05, "loss": 0.0046, "step": 155200 }, { "epoch": 15.62, "learning_rate": 1.0987087662665187e-05, "loss": 0.0047, "step": 155250 }, { "epoch": 15.63, "learning_rate": 1.0974477958236659e-05, "loss": 0.0046, "step": 155300 }, { "epoch": 15.63, "learning_rate": 1.0961868253808132e-05, "loss": 0.0039, "step": 155350 }, { "epoch": 15.64, "learning_rate": 1.0949258549379602e-05, "loss": 0.0055, "step": 155400 }, { "epoch": 15.64, "learning_rate": 1.0936648844951075e-05, "loss": 0.0055, "step": 155450 }, { "epoch": 15.65, "learning_rate": 1.0924039140522547e-05, "loss": 0.0046, "step": 155500 }, { "epoch": 15.65, "eval_cer": 0.05987583982465644, "eval_loss": 0.5481994152069092, "eval_runtime": 1309.4009, "eval_samples_per_second": 6.228, "eval_steps_per_second": 0.779, "step": 155500 }, { "epoch": 15.65, "learning_rate": 1.0911429436094018e-05, "loss": 0.0034, "step": 155550 }, { "epoch": 15.66, "learning_rate": 1.089881973166549e-05, "loss": 0.0038, "step": 155600 }, { "epoch": 15.66, "learning_rate": 1.0886210027236963e-05, "loss": 0.0039, "step": 155650 }, { "epoch": 15.67, "learning_rate": 1.0873600322808432e-05, "loss": 0.0048, "step": 155700 }, { "epoch": 15.67, "learning_rate": 1.0860990618379906e-05, "loss": 0.0043, "step": 155750 }, { "epoch": 15.68, "learning_rate": 1.0848380913951379e-05, "loss": 0.0032, "step": 155800 }, { "epoch": 15.68, "learning_rate": 1.0835771209522849e-05, "loss": 0.0046, "step": 155850 }, { "epoch": 15.69, "learning_rate": 1.0823161505094322e-05, "loss": 0.0047, "step": 155900 }, { "epoch": 15.69, "learning_rate": 1.0810551800665793e-05, "loss": 0.0036, "step": 155950 }, { "epoch": 15.7, "learning_rate": 1.0797942096237265e-05, "loss": 0.0046, "step": 156000 }, { "epoch": 15.7, "eval_cer": 0.05933100455170954, "eval_loss": 0.5516968369483948, "eval_runtime": 1350.1094, "eval_samples_per_second": 6.04, "eval_steps_per_second": 0.755, "step": 156000 }, { "epoch": 15.7, "learning_rate": 1.0785332391808736e-05, "loss": 0.0045, "step": 156050 }, { "epoch": 15.71, "learning_rate": 1.077272268738021e-05, "loss": 0.005, "step": 156100 }, { "epoch": 15.71, "learning_rate": 1.076011298295168e-05, "loss": 0.0049, "step": 156150 }, { "epoch": 15.72, "learning_rate": 1.0747503278523152e-05, "loss": 0.0035, "step": 156200 }, { "epoch": 15.72, "learning_rate": 1.0734893574094624e-05, "loss": 0.0038, "step": 156250 }, { "epoch": 15.73, "learning_rate": 1.0722283869666095e-05, "loss": 0.0042, "step": 156300 }, { "epoch": 15.73, "learning_rate": 1.0709674165237567e-05, "loss": 0.0046, "step": 156350 }, { "epoch": 15.74, "learning_rate": 1.069706446080904e-05, "loss": 0.0044, "step": 156400 }, { "epoch": 15.74, "learning_rate": 1.068445475638051e-05, "loss": 0.0036, "step": 156450 }, { "epoch": 15.75, "learning_rate": 1.0671845051951983e-05, "loss": 0.0043, "step": 156500 }, { "epoch": 15.75, "eval_cer": 0.059293644418707464, "eval_loss": 0.5518063306808472, "eval_runtime": 1625.9556, "eval_samples_per_second": 5.016, "eval_steps_per_second": 0.627, "step": 156500 }, { "epoch": 15.75, "learning_rate": 1.0659235347523454e-05, "loss": 0.0055, "step": 156550 }, { "epoch": 15.76, "learning_rate": 1.0646625643094926e-05, "loss": 0.004, "step": 156600 }, { "epoch": 15.76, "learning_rate": 1.0634015938666399e-05, "loss": 0.0049, "step": 156650 }, { "epoch": 15.77, "learning_rate": 1.062140623423787e-05, "loss": 0.004, "step": 156700 }, { "epoch": 15.77, "learning_rate": 1.0608796529809342e-05, "loss": 0.0039, "step": 156750 }, { "epoch": 15.78, "learning_rate": 1.0596186825380814e-05, "loss": 0.0075, "step": 156800 }, { "epoch": 15.78, "learning_rate": 1.0583577120952285e-05, "loss": 0.0043, "step": 156850 }, { "epoch": 15.79, "learning_rate": 1.0570967416523757e-05, "loss": 0.0052, "step": 156900 }, { "epoch": 15.79, "learning_rate": 1.055835771209523e-05, "loss": 0.006, "step": 156950 }, { "epoch": 15.8, "learning_rate": 1.0545748007666701e-05, "loss": 0.0035, "step": 157000 }, { "epoch": 15.8, "eval_cer": 0.05891070305543621, "eval_loss": 0.5524563789367676, "eval_runtime": 1638.54, "eval_samples_per_second": 4.977, "eval_steps_per_second": 0.623, "step": 157000 }, { "epoch": 15.8, "learning_rate": 1.0533138303238173e-05, "loss": 0.0035, "step": 157050 }, { "epoch": 15.81, "learning_rate": 1.0520528598809644e-05, "loss": 0.0047, "step": 157100 }, { "epoch": 15.81, "learning_rate": 1.0507918894381116e-05, "loss": 0.0053, "step": 157150 }, { "epoch": 15.82, "learning_rate": 1.0495309189952587e-05, "loss": 0.0032, "step": 157200 }, { "epoch": 15.82, "learning_rate": 1.048269948552406e-05, "loss": 0.0059, "step": 157250 }, { "epoch": 15.83, "learning_rate": 1.0470089781095532e-05, "loss": 0.0038, "step": 157300 }, { "epoch": 15.83, "learning_rate": 1.0457480076667003e-05, "loss": 0.0053, "step": 157350 }, { "epoch": 15.84, "learning_rate": 1.0444870372238476e-05, "loss": 0.0041, "step": 157400 }, { "epoch": 15.84, "learning_rate": 1.0432260667809946e-05, "loss": 0.0034, "step": 157450 }, { "epoch": 15.85, "learning_rate": 1.041965096338142e-05, "loss": 0.0042, "step": 157500 }, { "epoch": 15.85, "eval_cer": 0.05873012907925952, "eval_loss": 0.5525900721549988, "eval_runtime": 1397.614, "eval_samples_per_second": 5.835, "eval_steps_per_second": 0.73, "step": 157500 }, { "epoch": 15.85, "learning_rate": 1.0407041258952891e-05, "loss": 0.0046, "step": 157550 }, { "epoch": 15.86, "learning_rate": 1.0394431554524362e-05, "loss": 0.004, "step": 157600 }, { "epoch": 15.86, "learning_rate": 1.0381821850095834e-05, "loss": 0.0038, "step": 157650 }, { "epoch": 15.87, "learning_rate": 1.0369212145667307e-05, "loss": 0.0052, "step": 157700 }, { "epoch": 15.87, "learning_rate": 1.0356602441238777e-05, "loss": 0.0049, "step": 157750 }, { "epoch": 15.88, "learning_rate": 1.034399273681025e-05, "loss": 0.0045, "step": 157800 }, { "epoch": 15.88, "learning_rate": 1.0331383032381721e-05, "loss": 0.0042, "step": 157850 }, { "epoch": 15.89, "learning_rate": 1.0318773327953193e-05, "loss": 0.0046, "step": 157900 }, { "epoch": 15.89, "learning_rate": 1.0306163623524664e-05, "loss": 0.0039, "step": 157950 }, { "epoch": 15.9, "learning_rate": 1.0293806113184707e-05, "loss": 0.0047, "step": 158000 }, { "epoch": 15.9, "eval_cer": 0.05869276894625745, "eval_loss": 0.5514620542526245, "eval_runtime": 1313.2601, "eval_samples_per_second": 6.21, "eval_steps_per_second": 0.777, "step": 158000 }, { "epoch": 15.9, "learning_rate": 1.0281196408756179e-05, "loss": 0.0042, "step": 158050 }, { "epoch": 15.91, "learning_rate": 1.026858670432765e-05, "loss": 0.0046, "step": 158100 }, { "epoch": 15.91, "learning_rate": 1.0255976999899122e-05, "loss": 0.0051, "step": 158150 }, { "epoch": 15.92, "learning_rate": 1.0243367295470595e-05, "loss": 0.0055, "step": 158200 }, { "epoch": 15.92, "learning_rate": 1.0230757591042067e-05, "loss": 0.0033, "step": 158250 }, { "epoch": 15.93, "learning_rate": 1.0218147886613538e-05, "loss": 0.0049, "step": 158300 }, { "epoch": 15.93, "learning_rate": 1.0205538182185011e-05, "loss": 0.0036, "step": 158350 }, { "epoch": 15.94, "learning_rate": 1.0192928477756481e-05, "loss": 0.0055, "step": 158400 }, { "epoch": 15.94, "learning_rate": 1.0180318773327954e-05, "loss": 0.0041, "step": 158450 }, { "epoch": 15.95, "learning_rate": 1.0167709068899426e-05, "loss": 0.005, "step": 158500 }, { "epoch": 15.95, "eval_cer": 0.05883598278943206, "eval_loss": 0.5509193539619446, "eval_runtime": 1307.3147, "eval_samples_per_second": 6.238, "eval_steps_per_second": 0.78, "step": 158500 }, { "epoch": 15.95, "learning_rate": 1.0155099364470897e-05, "loss": 0.0029, "step": 158550 }, { "epoch": 15.96, "learning_rate": 1.0142489660042369e-05, "loss": 0.0037, "step": 158600 }, { "epoch": 15.96, "learning_rate": 1.0129879955613842e-05, "loss": 0.0029, "step": 158650 }, { "epoch": 15.97, "learning_rate": 1.0117270251185312e-05, "loss": 0.0031, "step": 158700 }, { "epoch": 15.97, "learning_rate": 1.0104660546756785e-05, "loss": 0.0048, "step": 158750 }, { "epoch": 15.98, "learning_rate": 1.0092050842328256e-05, "loss": 0.0035, "step": 158800 }, { "epoch": 15.98, "learning_rate": 1.0079441137899728e-05, "loss": 0.0043, "step": 158850 }, { "epoch": 15.99, "learning_rate": 1.006708362755977e-05, "loss": 0.0037, "step": 158900 }, { "epoch": 15.99, "learning_rate": 1.0054473923131242e-05, "loss": 0.0045, "step": 158950 }, { "epoch": 16.0, "learning_rate": 1.0041864218702714e-05, "loss": 0.0049, "step": 159000 }, { "epoch": 16.0, "eval_cer": 0.05906637027627818, "eval_loss": 0.5534919500350952, "eval_runtime": 1316.8673, "eval_samples_per_second": 6.193, "eval_steps_per_second": 0.775, "step": 159000 }, { "epoch": 16.0, "learning_rate": 1.0029254514274187e-05, "loss": 0.0032, "step": 159050 }, { "epoch": 16.01, "learning_rate": 1.0016644809845657e-05, "loss": 0.0027, "step": 159100 }, { "epoch": 16.01, "learning_rate": 1.000403510541713e-05, "loss": 0.0025, "step": 159150 }, { "epoch": 16.02, "learning_rate": 9.991425400988601e-06, "loss": 0.002, "step": 159200 }, { "epoch": 16.02, "learning_rate": 9.978815696560073e-06, "loss": 0.0027, "step": 159250 }, { "epoch": 16.03, "learning_rate": 9.966205992131544e-06, "loss": 0.0029, "step": 159300 }, { "epoch": 16.03, "learning_rate": 9.953596287703018e-06, "loss": 0.0037, "step": 159350 }, { "epoch": 16.04, "learning_rate": 9.940986583274489e-06, "loss": 0.0019, "step": 159400 }, { "epoch": 16.04, "learning_rate": 9.92837687884596e-06, "loss": 0.0035, "step": 159450 }, { "epoch": 16.05, "learning_rate": 9.915767174417432e-06, "loss": 0.002, "step": 159500 }, { "epoch": 16.05, "eval_cer": 0.058443701392910295, "eval_loss": 0.5554344058036804, "eval_runtime": 1676.5137, "eval_samples_per_second": 4.864, "eval_steps_per_second": 0.608, "step": 159500 }, { "epoch": 16.05, "learning_rate": 9.903157469988903e-06, "loss": 0.0034, "step": 159550 }, { "epoch": 16.06, "learning_rate": 9.890547765560377e-06, "loss": 0.003, "step": 159600 }, { "epoch": 16.06, "learning_rate": 9.877938061131848e-06, "loss": 0.0051, "step": 159650 }, { "epoch": 16.07, "learning_rate": 9.86532835670332e-06, "loss": 0.0032, "step": 159700 }, { "epoch": 16.07, "learning_rate": 9.852718652274791e-06, "loss": 0.0025, "step": 159750 }, { "epoch": 16.08, "learning_rate": 9.840108947846263e-06, "loss": 0.0032, "step": 159800 }, { "epoch": 16.08, "learning_rate": 9.827499243417734e-06, "loss": 0.0024, "step": 159850 }, { "epoch": 16.09, "learning_rate": 9.814889538989207e-06, "loss": 0.0028, "step": 159900 }, { "epoch": 16.09, "learning_rate": 9.802279834560677e-06, "loss": 0.002, "step": 159950 }, { "epoch": 16.1, "learning_rate": 9.78967013013215e-06, "loss": 0.0047, "step": 160000 }, { "epoch": 16.1, "eval_cer": 0.05886088954476678, "eval_loss": 0.5534769296646118, "eval_runtime": 1628.5356, "eval_samples_per_second": 5.008, "eval_steps_per_second": 0.626, "step": 160000 }, { "epoch": 16.1, "learning_rate": 9.777060425703622e-06, "loss": 0.0027, "step": 160050 }, { "epoch": 16.11, "learning_rate": 9.764450721275093e-06, "loss": 0.0037, "step": 160100 }, { "epoch": 16.11, "learning_rate": 9.751841016846566e-06, "loss": 0.0033, "step": 160150 }, { "epoch": 16.12, "learning_rate": 9.739231312418038e-06, "loss": 0.0028, "step": 160200 }, { "epoch": 16.12, "learning_rate": 9.72662160798951e-06, "loss": 0.0022, "step": 160250 }, { "epoch": 16.13, "learning_rate": 9.71401190356098e-06, "loss": 0.0028, "step": 160300 }, { "epoch": 16.14, "learning_rate": 9.701402199132454e-06, "loss": 0.0031, "step": 160350 }, { "epoch": 16.14, "learning_rate": 9.688792494703924e-06, "loss": 0.0031, "step": 160400 }, { "epoch": 16.15, "learning_rate": 9.676182790275397e-06, "loss": 0.0037, "step": 160450 }, { "epoch": 16.15, "learning_rate": 9.663573085846868e-06, "loss": 0.0038, "step": 160500 }, { "epoch": 16.15, "eval_cer": 0.0582786941388178, "eval_loss": 0.553225040435791, "eval_runtime": 1608.2979, "eval_samples_per_second": 5.071, "eval_steps_per_second": 0.634, "step": 160500 }, { "epoch": 16.16, "learning_rate": 9.65096338141834e-06, "loss": 0.0043, "step": 160550 }, { "epoch": 16.16, "learning_rate": 9.638353676989811e-06, "loss": 0.0032, "step": 160600 }, { "epoch": 16.17, "learning_rate": 9.625743972561285e-06, "loss": 0.0035, "step": 160650 }, { "epoch": 16.17, "learning_rate": 9.613134268132754e-06, "loss": 0.0034, "step": 160700 }, { "epoch": 16.18, "learning_rate": 9.600524563704228e-06, "loss": 0.0035, "step": 160750 }, { "epoch": 16.18, "learning_rate": 9.587914859275699e-06, "loss": 0.0036, "step": 160800 }, { "epoch": 16.19, "learning_rate": 9.57530515484717e-06, "loss": 0.0029, "step": 160850 }, { "epoch": 16.19, "learning_rate": 9.562695450418642e-06, "loss": 0.0028, "step": 160900 }, { "epoch": 16.2, "learning_rate": 9.550085745990115e-06, "loss": 0.0028, "step": 160950 }, { "epoch": 16.2, "learning_rate": 9.537476041561587e-06, "loss": 0.0027, "step": 161000 }, { "epoch": 16.2, "eval_cer": 0.058322280960653554, "eval_loss": 0.5514861941337585, "eval_runtime": 1481.9491, "eval_samples_per_second": 5.503, "eval_steps_per_second": 0.688, "step": 161000 }, { "epoch": 16.21, "learning_rate": 9.524866337133058e-06, "loss": 0.0023, "step": 161050 }, { "epoch": 16.21, "learning_rate": 9.51225663270453e-06, "loss": 0.0031, "step": 161100 }, { "epoch": 16.22, "learning_rate": 9.499646928276001e-06, "loss": 0.0041, "step": 161150 }, { "epoch": 16.22, "learning_rate": 9.487037223847474e-06, "loss": 0.002, "step": 161200 }, { "epoch": 16.23, "learning_rate": 9.474427519418946e-06, "loss": 0.0023, "step": 161250 }, { "epoch": 16.23, "learning_rate": 9.461817814990417e-06, "loss": 0.0024, "step": 161300 }, { "epoch": 16.24, "learning_rate": 9.449208110561889e-06, "loss": 0.0029, "step": 161350 }, { "epoch": 16.24, "learning_rate": 9.43659840613336e-06, "loss": 0.0026, "step": 161400 }, { "epoch": 16.25, "learning_rate": 9.423988701704832e-06, "loss": 0.0035, "step": 161450 }, { "epoch": 16.25, "learning_rate": 9.411378997276305e-06, "loss": 0.0058, "step": 161500 }, { "epoch": 16.25, "eval_cer": 0.05813548029564319, "eval_loss": 0.5503236651420593, "eval_runtime": 1305.3358, "eval_samples_per_second": 6.247, "eval_steps_per_second": 0.781, "step": 161500 }, { "epoch": 16.26, "learning_rate": 9.398769292847776e-06, "loss": 0.0031, "step": 161550 }, { "epoch": 16.26, "learning_rate": 9.386159588419248e-06, "loss": 0.0027, "step": 161600 }, { "epoch": 16.27, "learning_rate": 9.37354988399072e-06, "loss": 0.0028, "step": 161650 }, { "epoch": 16.27, "learning_rate": 9.36094017956219e-06, "loss": 0.0033, "step": 161700 }, { "epoch": 16.28, "learning_rate": 9.348330475133664e-06, "loss": 0.0042, "step": 161750 }, { "epoch": 16.28, "learning_rate": 9.335720770705135e-06, "loss": 0.0033, "step": 161800 }, { "epoch": 16.29, "learning_rate": 9.323111066276607e-06, "loss": 0.005, "step": 161850 }, { "epoch": 16.29, "learning_rate": 9.310501361848078e-06, "loss": 0.0028, "step": 161900 }, { "epoch": 16.3, "learning_rate": 9.297891657419552e-06, "loss": 0.0021, "step": 161950 }, { "epoch": 16.3, "learning_rate": 9.285281952991021e-06, "loss": 0.0024, "step": 162000 }, { "epoch": 16.3, "eval_cer": 0.057986039763634895, "eval_loss": 0.5542289614677429, "eval_runtime": 1524.4829, "eval_samples_per_second": 5.349, "eval_steps_per_second": 0.669, "step": 162000 }, { "epoch": 16.31, "learning_rate": 9.272672248562495e-06, "loss": 0.0023, "step": 162050 }, { "epoch": 16.31, "learning_rate": 9.260062544133966e-06, "loss": 0.0033, "step": 162100 }, { "epoch": 16.32, "learning_rate": 9.247452839705438e-06, "loss": 0.0029, "step": 162150 }, { "epoch": 16.32, "learning_rate": 9.234843135276909e-06, "loss": 0.0031, "step": 162200 }, { "epoch": 16.33, "learning_rate": 9.222233430848382e-06, "loss": 0.0021, "step": 162250 }, { "epoch": 16.33, "learning_rate": 9.209623726419852e-06, "loss": 0.0024, "step": 162300 }, { "epoch": 16.34, "learning_rate": 9.197014021991325e-06, "loss": 0.0024, "step": 162350 }, { "epoch": 16.34, "learning_rate": 9.184404317562797e-06, "loss": 0.0029, "step": 162400 }, { "epoch": 16.35, "learning_rate": 9.171794613134268e-06, "loss": 0.0033, "step": 162450 }, { "epoch": 16.35, "learning_rate": 9.159184908705741e-06, "loss": 0.0025, "step": 162500 }, { "epoch": 16.35, "eval_cer": 0.058147933673310545, "eval_loss": 0.5534273386001587, "eval_runtime": 1614.3844, "eval_samples_per_second": 5.051, "eval_steps_per_second": 0.632, "step": 162500 }, { "epoch": 16.36, "learning_rate": 9.146575204277213e-06, "loss": 0.0034, "step": 162550 }, { "epoch": 16.36, "learning_rate": 9.133965499848684e-06, "loss": 0.0053, "step": 162600 }, { "epoch": 16.37, "learning_rate": 9.121355795420156e-06, "loss": 0.0041, "step": 162650 }, { "epoch": 16.37, "learning_rate": 9.108998285080197e-06, "loss": 0.0028, "step": 162700 }, { "epoch": 16.38, "learning_rate": 9.09638858065167e-06, "loss": 0.0036, "step": 162750 }, { "epoch": 16.38, "learning_rate": 9.083778876223142e-06, "loss": 0.0023, "step": 162800 }, { "epoch": 16.39, "learning_rate": 9.071169171794613e-06, "loss": 0.0031, "step": 162850 }, { "epoch": 16.39, "learning_rate": 9.058559467366086e-06, "loss": 0.0032, "step": 162900 }, { "epoch": 16.4, "learning_rate": 9.045949762937556e-06, "loss": 0.0022, "step": 162950 }, { "epoch": 16.4, "learning_rate": 9.03334005850903e-06, "loss": 0.0031, "step": 163000 }, { "epoch": 16.4, "eval_cer": 0.05828803417206832, "eval_loss": 0.5522481799125671, "eval_runtime": 1440.8838, "eval_samples_per_second": 5.66, "eval_steps_per_second": 0.708, "step": 163000 }, { "epoch": 16.41, "learning_rate": 9.0207303540805e-06, "loss": 0.0025, "step": 163050 }, { "epoch": 16.41, "learning_rate": 9.008120649651972e-06, "loss": 0.0023, "step": 163100 }, { "epoch": 16.42, "learning_rate": 8.995510945223444e-06, "loss": 0.0029, "step": 163150 }, { "epoch": 16.42, "learning_rate": 8.982901240794917e-06, "loss": 0.003, "step": 163200 }, { "epoch": 16.43, "learning_rate": 8.970291536366387e-06, "loss": 0.0038, "step": 163250 }, { "epoch": 16.43, "learning_rate": 8.95768183193786e-06, "loss": 0.0025, "step": 163300 }, { "epoch": 16.44, "learning_rate": 8.945072127509331e-06, "loss": 0.0044, "step": 163350 }, { "epoch": 16.44, "learning_rate": 8.932462423080803e-06, "loss": 0.0018, "step": 163400 }, { "epoch": 16.45, "learning_rate": 8.919852718652274e-06, "loss": 0.0033, "step": 163450 }, { "epoch": 16.45, "learning_rate": 8.907243014223748e-06, "loss": 0.0023, "step": 163500 }, { "epoch": 16.45, "eval_cer": 0.05813548029564319, "eval_loss": 0.5540639758110046, "eval_runtime": 1307.1658, "eval_samples_per_second": 6.239, "eval_steps_per_second": 0.78, "step": 163500 }, { "epoch": 16.46, "learning_rate": 8.894633309795219e-06, "loss": 0.0022, "step": 163550 }, { "epoch": 16.46, "learning_rate": 8.88202360536669e-06, "loss": 0.0029, "step": 163600 }, { "epoch": 16.47, "learning_rate": 8.869413900938164e-06, "loss": 0.0033, "step": 163650 }, { "epoch": 16.47, "learning_rate": 8.856804196509633e-06, "loss": 0.0031, "step": 163700 }, { "epoch": 16.48, "learning_rate": 8.844194492081107e-06, "loss": 0.0044, "step": 163750 }, { "epoch": 16.48, "learning_rate": 8.831584787652578e-06, "loss": 0.0035, "step": 163800 }, { "epoch": 16.49, "learning_rate": 8.81897508322405e-06, "loss": 0.0026, "step": 163850 }, { "epoch": 16.49, "learning_rate": 8.806365378795521e-06, "loss": 0.0032, "step": 163900 }, { "epoch": 16.5, "learning_rate": 8.793755674366994e-06, "loss": 0.0022, "step": 163950 }, { "epoch": 16.5, "learning_rate": 8.781145969938464e-06, "loss": 0.003, "step": 164000 }, { "epoch": 16.5, "eval_cer": 0.05794556628621598, "eval_loss": 0.5508033633232117, "eval_runtime": 1305.2487, "eval_samples_per_second": 6.248, "eval_steps_per_second": 0.781, "step": 164000 }, { "epoch": 16.51, "learning_rate": 8.768536265509937e-06, "loss": 0.0025, "step": 164050 }, { "epoch": 16.51, "learning_rate": 8.755926561081409e-06, "loss": 0.0024, "step": 164100 }, { "epoch": 16.52, "learning_rate": 8.74331685665288e-06, "loss": 0.0025, "step": 164150 }, { "epoch": 16.52, "learning_rate": 8.730707152224352e-06, "loss": 0.0027, "step": 164200 }, { "epoch": 16.53, "learning_rate": 8.718097447795825e-06, "loss": 0.0036, "step": 164250 }, { "epoch": 16.53, "learning_rate": 8.705487743367296e-06, "loss": 0.0023, "step": 164300 }, { "epoch": 16.54, "learning_rate": 8.692878038938768e-06, "loss": 0.0043, "step": 164350 }, { "epoch": 16.54, "learning_rate": 8.68026833451024e-06, "loss": 0.0014, "step": 164400 }, { "epoch": 16.55, "learning_rate": 8.66765863008171e-06, "loss": 0.0031, "step": 164450 }, { "epoch": 16.55, "learning_rate": 8.655048925653184e-06, "loss": 0.0024, "step": 164500 }, { "epoch": 16.55, "eval_cer": 0.05811057354030847, "eval_loss": 0.5504255890846252, "eval_runtime": 1307.1727, "eval_samples_per_second": 6.239, "eval_steps_per_second": 0.78, "step": 164500 }, { "epoch": 16.56, "learning_rate": 8.642691415313225e-06, "loss": 0.0026, "step": 164550 }, { "epoch": 16.56, "learning_rate": 8.630081710884698e-06, "loss": 0.003, "step": 164600 }, { "epoch": 16.57, "learning_rate": 8.617472006456168e-06, "loss": 0.0025, "step": 164650 }, { "epoch": 16.57, "learning_rate": 8.604862302027641e-06, "loss": 0.0024, "step": 164700 }, { "epoch": 16.58, "learning_rate": 8.592252597599113e-06, "loss": 0.0036, "step": 164750 }, { "epoch": 16.58, "learning_rate": 8.579642893170584e-06, "loss": 0.0026, "step": 164800 }, { "epoch": 16.59, "learning_rate": 8.567033188742056e-06, "loss": 0.0039, "step": 164850 }, { "epoch": 16.59, "learning_rate": 8.554423484313529e-06, "loss": 0.0031, "step": 164900 }, { "epoch": 16.6, "learning_rate": 8.541813779884999e-06, "loss": 0.0035, "step": 164950 }, { "epoch": 16.6, "learning_rate": 8.529204075456472e-06, "loss": 0.0028, "step": 165000 }, { "epoch": 16.6, "eval_cer": 0.05824133400581573, "eval_loss": 0.551003634929657, "eval_runtime": 1306.5317, "eval_samples_per_second": 6.242, "eval_steps_per_second": 0.781, "step": 165000 }, { "epoch": 16.61, "learning_rate": 8.516594371027944e-06, "loss": 0.005, "step": 165050 }, { "epoch": 16.61, "learning_rate": 8.503984666599415e-06, "loss": 0.0033, "step": 165100 }, { "epoch": 16.62, "learning_rate": 8.491374962170887e-06, "loss": 0.0016, "step": 165150 }, { "epoch": 16.62, "learning_rate": 8.47876525774236e-06, "loss": 0.0021, "step": 165200 }, { "epoch": 16.63, "learning_rate": 8.46615555331383e-06, "loss": 0.002, "step": 165250 }, { "epoch": 16.63, "learning_rate": 8.453545848885303e-06, "loss": 0.003, "step": 165300 }, { "epoch": 16.64, "learning_rate": 8.440936144456776e-06, "loss": 0.0029, "step": 165350 }, { "epoch": 16.64, "learning_rate": 8.428326440028246e-06, "loss": 0.0023, "step": 165400 }, { "epoch": 16.65, "learning_rate": 8.415716735599719e-06, "loss": 0.0025, "step": 165450 }, { "epoch": 16.65, "learning_rate": 8.40310703117119e-06, "loss": 0.0021, "step": 165500 }, { "epoch": 16.65, "eval_cer": 0.05829737420531884, "eval_loss": 0.5551767945289612, "eval_runtime": 1306.837, "eval_samples_per_second": 6.24, "eval_steps_per_second": 0.781, "step": 165500 }, { "epoch": 16.66, "learning_rate": 8.390497326742662e-06, "loss": 0.0027, "step": 165550 }, { "epoch": 16.66, "learning_rate": 8.377887622314133e-06, "loss": 0.0041, "step": 165600 }, { "epoch": 16.67, "learning_rate": 8.365277917885605e-06, "loss": 0.0027, "step": 165650 }, { "epoch": 16.67, "learning_rate": 8.352668213457076e-06, "loss": 0.003, "step": 165700 }, { "epoch": 16.68, "learning_rate": 8.34005850902855e-06, "loss": 0.0032, "step": 165750 }, { "epoch": 16.68, "learning_rate": 8.327448804600021e-06, "loss": 0.0028, "step": 165800 }, { "epoch": 16.69, "learning_rate": 8.314839100171492e-06, "loss": 0.0046, "step": 165850 }, { "epoch": 16.69, "learning_rate": 8.302229395742964e-06, "loss": 0.0028, "step": 165900 }, { "epoch": 16.7, "learning_rate": 8.289619691314435e-06, "loss": 0.0033, "step": 165950 }, { "epoch": 16.7, "learning_rate": 8.277009986885907e-06, "loss": 0.0032, "step": 166000 }, { "epoch": 16.7, "eval_cer": 0.05839077453782402, "eval_loss": 0.5505055785179138, "eval_runtime": 1305.1091, "eval_samples_per_second": 6.249, "eval_steps_per_second": 0.782, "step": 166000 }, { "epoch": 16.71, "learning_rate": 8.26440028245738e-06, "loss": 0.0037, "step": 166050 }, { "epoch": 16.71, "learning_rate": 8.251790578028851e-06, "loss": 0.0023, "step": 166100 }, { "epoch": 16.72, "learning_rate": 8.239180873600323e-06, "loss": 0.0037, "step": 166150 }, { "epoch": 16.72, "learning_rate": 8.226571169171796e-06, "loss": 0.0028, "step": 166200 }, { "epoch": 16.73, "learning_rate": 8.213961464743266e-06, "loss": 0.002, "step": 166250 }, { "epoch": 16.73, "learning_rate": 8.201351760314739e-06, "loss": 0.0022, "step": 166300 }, { "epoch": 16.74, "learning_rate": 8.18874205588621e-06, "loss": 0.0022, "step": 166350 }, { "epoch": 16.74, "learning_rate": 8.176132351457682e-06, "loss": 0.0024, "step": 166400 }, { "epoch": 16.75, "learning_rate": 8.163522647029154e-06, "loss": 0.0026, "step": 166450 }, { "epoch": 16.75, "learning_rate": 8.150912942600627e-06, "loss": 0.0023, "step": 166500 }, { "epoch": 16.75, "eval_cer": 0.057914432842047586, "eval_loss": 0.5533706545829773, "eval_runtime": 1305.465, "eval_samples_per_second": 6.247, "eval_steps_per_second": 0.781, "step": 166500 }, { "epoch": 16.76, "learning_rate": 8.138303238172097e-06, "loss": 0.0024, "step": 166550 }, { "epoch": 16.76, "learning_rate": 8.12569353374357e-06, "loss": 0.0017, "step": 166600 }, { "epoch": 16.77, "learning_rate": 8.113083829315041e-06, "loss": 0.0018, "step": 166650 }, { "epoch": 16.77, "learning_rate": 8.100474124886513e-06, "loss": 0.005, "step": 166700 }, { "epoch": 16.78, "learning_rate": 8.087864420457984e-06, "loss": 0.003, "step": 166750 }, { "epoch": 16.78, "learning_rate": 8.075254716029457e-06, "loss": 0.0023, "step": 166800 }, { "epoch": 16.79, "learning_rate": 8.062645011600927e-06, "loss": 0.0025, "step": 166850 }, { "epoch": 16.79, "learning_rate": 8.0500353071724e-06, "loss": 0.0029, "step": 166900 }, { "epoch": 16.8, "learning_rate": 8.037425602743873e-06, "loss": 0.0027, "step": 166950 }, { "epoch": 16.8, "learning_rate": 8.024815898315343e-06, "loss": 0.0028, "step": 167000 }, { "epoch": 16.8, "eval_cer": 0.058088780129390595, "eval_loss": 0.5518389940261841, "eval_runtime": 1304.1385, "eval_samples_per_second": 6.253, "eval_steps_per_second": 0.782, "step": 167000 }, { "epoch": 16.81, "learning_rate": 8.012206193886816e-06, "loss": 0.0039, "step": 167050 }, { "epoch": 16.81, "learning_rate": 7.999596489458288e-06, "loss": 0.0031, "step": 167100 }, { "epoch": 16.82, "learning_rate": 7.98698678502976e-06, "loss": 0.0029, "step": 167150 }, { "epoch": 16.82, "learning_rate": 7.974377080601231e-06, "loss": 0.0036, "step": 167200 }, { "epoch": 16.83, "learning_rate": 7.961767376172704e-06, "loss": 0.0024, "step": 167250 }, { "epoch": 16.83, "learning_rate": 7.949157671744174e-06, "loss": 0.0023, "step": 167300 }, { "epoch": 16.84, "learning_rate": 7.936547967315647e-06, "loss": 0.0022, "step": 167350 }, { "epoch": 16.84, "learning_rate": 7.923938262887119e-06, "loss": 0.0026, "step": 167400 }, { "epoch": 16.85, "learning_rate": 7.91132855845859e-06, "loss": 0.0032, "step": 167450 }, { "epoch": 16.85, "learning_rate": 7.898718854030061e-06, "loss": 0.0028, "step": 167500 }, { "epoch": 16.85, "eval_cer": 0.058605595302585946, "eval_loss": 0.5523379445075989, "eval_runtime": 1304.8929, "eval_samples_per_second": 6.25, "eval_steps_per_second": 0.782, "step": 167500 }, { "epoch": 16.86, "learning_rate": 7.886109149601533e-06, "loss": 0.0029, "step": 167550 }, { "epoch": 16.86, "learning_rate": 7.873499445173004e-06, "loss": 0.0034, "step": 167600 }, { "epoch": 16.87, "learning_rate": 7.860889740744478e-06, "loss": 0.003, "step": 167650 }, { "epoch": 16.87, "learning_rate": 7.848280036315949e-06, "loss": 0.0026, "step": 167700 }, { "epoch": 16.88, "learning_rate": 7.83567033188742e-06, "loss": 0.0025, "step": 167750 }, { "epoch": 16.88, "learning_rate": 7.823060627458894e-06, "loss": 0.0029, "step": 167800 }, { "epoch": 16.89, "learning_rate": 7.810450923030364e-06, "loss": 0.0028, "step": 167850 }, { "epoch": 16.89, "learning_rate": 7.797841218601837e-06, "loss": 0.0022, "step": 167900 }, { "epoch": 16.9, "learning_rate": 7.785231514173308e-06, "loss": 0.0027, "step": 167950 }, { "epoch": 16.9, "learning_rate": 7.77262180974478e-06, "loss": 0.0022, "step": 168000 }, { "epoch": 16.9, "eval_cer": 0.05792999956413178, "eval_loss": 0.5529240369796753, "eval_runtime": 1302.3306, "eval_samples_per_second": 6.262, "eval_steps_per_second": 0.783, "step": 168000 }, { "epoch": 16.91, "learning_rate": 7.760012105316251e-06, "loss": 0.0032, "step": 168050 }, { "epoch": 16.91, "learning_rate": 7.747402400887724e-06, "loss": 0.0028, "step": 168100 }, { "epoch": 16.92, "learning_rate": 7.734792696459194e-06, "loss": 0.0028, "step": 168150 }, { "epoch": 16.92, "learning_rate": 7.722182992030667e-06, "loss": 0.0046, "step": 168200 }, { "epoch": 16.93, "learning_rate": 7.709573287602139e-06, "loss": 0.0038, "step": 168250 }, { "epoch": 16.93, "learning_rate": 7.69696358317361e-06, "loss": 0.0029, "step": 168300 }, { "epoch": 16.94, "learning_rate": 7.684353878745082e-06, "loss": 0.0037, "step": 168350 }, { "epoch": 16.95, "learning_rate": 7.671744174316555e-06, "loss": 0.0026, "step": 168400 }, { "epoch": 16.95, "learning_rate": 7.659134469888026e-06, "loss": 0.0029, "step": 168450 }, { "epoch": 16.96, "learning_rate": 7.646524765459498e-06, "loss": 0.0017, "step": 168500 }, { "epoch": 16.96, "eval_cer": 0.05816350039539474, "eval_loss": 0.5528781414031982, "eval_runtime": 1308.6326, "eval_samples_per_second": 6.232, "eval_steps_per_second": 0.779, "step": 168500 }, { "epoch": 16.96, "learning_rate": 7.633915061030971e-06, "loss": 0.0036, "step": 168550 }, { "epoch": 16.97, "learning_rate": 7.621305356602442e-06, "loss": 0.0027, "step": 168600 }, { "epoch": 16.97, "learning_rate": 7.608695652173914e-06, "loss": 0.0108, "step": 168650 }, { "epoch": 16.98, "learning_rate": 7.596085947745385e-06, "loss": 0.003, "step": 168700 }, { "epoch": 16.98, "learning_rate": 7.583476243316857e-06, "loss": 0.0022, "step": 168750 }, { "epoch": 16.99, "learning_rate": 7.5708665388883285e-06, "loss": 0.0028, "step": 168800 }, { "epoch": 16.99, "learning_rate": 7.558256834459801e-06, "loss": 0.0037, "step": 168850 }, { "epoch": 17.0, "learning_rate": 7.545647130031272e-06, "loss": 0.002, "step": 168900 }, { "epoch": 17.0, "learning_rate": 7.533037425602745e-06, "loss": 0.002, "step": 168950 }, { "epoch": 17.01, "learning_rate": 7.520427721174215e-06, "loss": 0.0022, "step": 169000 }, { "epoch": 17.01, "eval_cer": 0.05805141999638852, "eval_loss": 0.5554443001747131, "eval_runtime": 1304.9804, "eval_samples_per_second": 6.249, "eval_steps_per_second": 0.782, "step": 169000 }, { "epoch": 17.01, "learning_rate": 7.507818016745688e-06, "loss": 0.0024, "step": 169050 }, { "epoch": 17.02, "learning_rate": 7.49546050640573e-06, "loss": 0.0016, "step": 169100 }, { "epoch": 17.02, "learning_rate": 7.482850801977202e-06, "loss": 0.0015, "step": 169150 }, { "epoch": 17.03, "learning_rate": 7.470241097548674e-06, "loss": 0.0014, "step": 169200 }, { "epoch": 17.03, "learning_rate": 7.457631393120146e-06, "loss": 0.0012, "step": 169250 }, { "epoch": 17.04, "learning_rate": 7.4450216886916166e-06, "loss": 0.0021, "step": 169300 }, { "epoch": 17.04, "learning_rate": 7.432411984263089e-06, "loss": 0.0017, "step": 169350 }, { "epoch": 17.05, "learning_rate": 7.41980227983456e-06, "loss": 0.0019, "step": 169400 }, { "epoch": 17.05, "learning_rate": 7.407192575406033e-06, "loss": 0.0021, "step": 169450 }, { "epoch": 17.06, "learning_rate": 7.394582870977505e-06, "loss": 0.0022, "step": 169500 }, { "epoch": 17.06, "eval_cer": 0.058070100062889554, "eval_loss": 0.5544934868812561, "eval_runtime": 1306.7702, "eval_samples_per_second": 6.241, "eval_steps_per_second": 0.781, "step": 169500 }, { "epoch": 17.06, "learning_rate": 7.3819731665489765e-06, "loss": 0.002, "step": 169550 }, { "epoch": 17.07, "learning_rate": 7.369363462120449e-06, "loss": 0.0019, "step": 169600 }, { "epoch": 17.07, "learning_rate": 7.3567537576919195e-06, "loss": 0.0033, "step": 169650 }, { "epoch": 17.08, "learning_rate": 7.344144053263393e-06, "loss": 0.0014, "step": 169700 }, { "epoch": 17.08, "learning_rate": 7.331534348834863e-06, "loss": 0.0023, "step": 169750 }, { "epoch": 17.09, "learning_rate": 7.318924644406336e-06, "loss": 0.0023, "step": 169800 }, { "epoch": 17.09, "learning_rate": 7.306314939977807e-06, "loss": 0.0014, "step": 169850 }, { "epoch": 17.1, "learning_rate": 7.2937052355492795e-06, "loss": 0.0018, "step": 169900 }, { "epoch": 17.1, "learning_rate": 7.28109553112075e-06, "loss": 0.003, "step": 169950 }, { "epoch": 17.11, "learning_rate": 7.268485826692223e-06, "loss": 0.0029, "step": 170000 }, { "epoch": 17.11, "eval_cer": 0.0575688516117784, "eval_loss": 0.5546350479125977, "eval_runtime": 1302.4251, "eval_samples_per_second": 6.261, "eval_steps_per_second": 0.783, "step": 170000 }, { "epoch": 17.11, "learning_rate": 7.255876122263694e-06, "loss": 0.0025, "step": 170050 }, { "epoch": 17.12, "learning_rate": 7.243266417835166e-06, "loss": 0.0014, "step": 170100 }, { "epoch": 17.12, "learning_rate": 7.230656713406638e-06, "loss": 0.0016, "step": 170150 }, { "epoch": 17.13, "learning_rate": 7.21804700897811e-06, "loss": 0.0017, "step": 170200 }, { "epoch": 17.13, "learning_rate": 7.205437304549582e-06, "loss": 0.0019, "step": 170250 }, { "epoch": 17.14, "learning_rate": 7.192827600121054e-06, "loss": 0.0018, "step": 170300 }, { "epoch": 17.14, "learning_rate": 7.180217895692526e-06, "loss": 0.002, "step": 170350 }, { "epoch": 17.15, "learning_rate": 7.167608191263997e-06, "loss": 0.0026, "step": 170400 }, { "epoch": 17.15, "learning_rate": 7.154998486835469e-06, "loss": 0.0019, "step": 170450 }, { "epoch": 17.16, "learning_rate": 7.142388782406941e-06, "loss": 0.0033, "step": 170500 }, { "epoch": 17.16, "eval_cer": 0.05771517879936986, "eval_loss": 0.5562267899513245, "eval_runtime": 1302.4639, "eval_samples_per_second": 6.261, "eval_steps_per_second": 0.783, "step": 170500 }, { "epoch": 17.16, "learning_rate": 7.129779077978413e-06, "loss": 0.0024, "step": 170550 }, { "epoch": 17.17, "learning_rate": 7.117169373549884e-06, "loss": 0.0015, "step": 170600 }, { "epoch": 17.17, "learning_rate": 7.104559669121357e-06, "loss": 0.0022, "step": 170650 }, { "epoch": 17.18, "learning_rate": 7.0919499646928274e-06, "loss": 0.0028, "step": 170700 }, { "epoch": 17.18, "learning_rate": 7.0793402602643e-06, "loss": 0.002, "step": 170750 }, { "epoch": 17.19, "learning_rate": 7.066730555835771e-06, "loss": 0.0019, "step": 170800 }, { "epoch": 17.19, "learning_rate": 7.0541208514072436e-06, "loss": 0.0019, "step": 170850 }, { "epoch": 17.2, "learning_rate": 7.041511146978714e-06, "loss": 0.0012, "step": 170900 }, { "epoch": 17.2, "learning_rate": 7.028901442550187e-06, "loss": 0.0025, "step": 170950 }, { "epoch": 17.21, "learning_rate": 7.016291738121658e-06, "loss": 0.0016, "step": 171000 }, { "epoch": 17.21, "eval_cer": 0.05781480582070872, "eval_loss": 0.5555543303489685, "eval_runtime": 1302.4601, "eval_samples_per_second": 6.261, "eval_steps_per_second": 0.783, "step": 171000 }, { "epoch": 17.21, "learning_rate": 7.00368203369313e-06, "loss": 0.0016, "step": 171050 }, { "epoch": 17.22, "learning_rate": 6.991072329264603e-06, "loss": 0.0015, "step": 171100 }, { "epoch": 17.22, "learning_rate": 6.978462624836074e-06, "loss": 0.0012, "step": 171150 }, { "epoch": 17.23, "learning_rate": 6.9658529204075465e-06, "loss": 0.0028, "step": 171200 }, { "epoch": 17.23, "learning_rate": 6.953243215979018e-06, "loss": 0.0025, "step": 171250 }, { "epoch": 17.24, "learning_rate": 6.94063351155049e-06, "loss": 0.0017, "step": 171300 }, { "epoch": 17.24, "learning_rate": 6.928023807121961e-06, "loss": 0.0021, "step": 171350 }, { "epoch": 17.25, "learning_rate": 6.915414102693433e-06, "loss": 0.0025, "step": 171400 }, { "epoch": 17.25, "learning_rate": 6.902804398264905e-06, "loss": 0.0015, "step": 171450 }, { "epoch": 17.26, "learning_rate": 6.890194693836377e-06, "loss": 0.002, "step": 171500 }, { "epoch": 17.26, "eval_cer": 0.057914432842047586, "eval_loss": 0.5549079775810242, "eval_runtime": 1330.5514, "eval_samples_per_second": 6.129, "eval_steps_per_second": 0.767, "step": 171500 }, { "epoch": 17.26, "learning_rate": 6.877584989407848e-06, "loss": 0.0015, "step": 171550 }, { "epoch": 17.27, "learning_rate": 6.864975284979321e-06, "loss": 0.0023, "step": 171600 }, { "epoch": 17.27, "learning_rate": 6.8523655805507915e-06, "loss": 0.0021, "step": 171650 }, { "epoch": 17.28, "learning_rate": 6.839755876122264e-06, "loss": 0.0017, "step": 171700 }, { "epoch": 17.28, "learning_rate": 6.827146171693735e-06, "loss": 0.0015, "step": 171750 }, { "epoch": 17.29, "learning_rate": 6.814536467265208e-06, "loss": 0.0089, "step": 171800 }, { "epoch": 17.29, "learning_rate": 6.80192676283668e-06, "loss": 0.0018, "step": 171850 }, { "epoch": 17.3, "learning_rate": 6.7893170584081515e-06, "loss": 0.0018, "step": 171900 }, { "epoch": 17.3, "learning_rate": 6.776707353979624e-06, "loss": 0.0017, "step": 171950 }, { "epoch": 17.31, "learning_rate": 6.7640976495510945e-06, "loss": 0.0014, "step": 172000 }, { "epoch": 17.31, "eval_cer": 0.057634231844532036, "eval_loss": 0.5557627081871033, "eval_runtime": 1701.0477, "eval_samples_per_second": 4.794, "eval_steps_per_second": 0.6, "step": 172000 }, { "epoch": 17.31, "learning_rate": 6.751487945122567e-06, "loss": 0.002, "step": 172050 }, { "epoch": 17.32, "learning_rate": 6.738878240694038e-06, "loss": 0.002, "step": 172100 }, { "epoch": 17.32, "learning_rate": 6.726268536265511e-06, "loss": 0.0025, "step": 172150 }, { "epoch": 17.33, "learning_rate": 6.713658831836982e-06, "loss": 0.0018, "step": 172200 }, { "epoch": 17.33, "learning_rate": 6.7010491274084544e-06, "loss": 0.003, "step": 172250 }, { "epoch": 17.34, "learning_rate": 6.688439422979925e-06, "loss": 0.0019, "step": 172300 }, { "epoch": 17.34, "learning_rate": 6.675829718551397e-06, "loss": 0.0025, "step": 172350 }, { "epoch": 17.35, "learning_rate": 6.663220014122869e-06, "loss": 0.003, "step": 172400 }, { "epoch": 17.35, "learning_rate": 6.650610309694341e-06, "loss": 0.0013, "step": 172450 }, { "epoch": 17.36, "learning_rate": 6.638000605265812e-06, "loss": 0.0026, "step": 172500 }, { "epoch": 17.36, "eval_cer": 0.05781480582070872, "eval_loss": 0.5572019815444946, "eval_runtime": 1690.6451, "eval_samples_per_second": 4.824, "eval_steps_per_second": 0.603, "step": 172500 }, { "epoch": 17.36, "learning_rate": 6.625390900837285e-06, "loss": 0.0022, "step": 172550 }, { "epoch": 17.37, "learning_rate": 6.612781196408757e-06, "loss": 0.0019, "step": 172600 }, { "epoch": 17.37, "learning_rate": 6.600171491980228e-06, "loss": 0.0024, "step": 172650 }, { "epoch": 17.38, "learning_rate": 6.5875617875517e-06, "loss": 0.0029, "step": 172700 }, { "epoch": 17.38, "learning_rate": 6.574952083123172e-06, "loss": 0.0022, "step": 172750 }, { "epoch": 17.39, "learning_rate": 6.562342378694644e-06, "loss": 0.0017, "step": 172800 }, { "epoch": 17.39, "learning_rate": 6.549732674266116e-06, "loss": 0.0015, "step": 172850 }, { "epoch": 17.4, "learning_rate": 6.537122969837588e-06, "loss": 0.0013, "step": 172900 }, { "epoch": 17.4, "learning_rate": 6.524513265409059e-06, "loss": 0.0014, "step": 172950 }, { "epoch": 17.41, "learning_rate": 6.511903560980531e-06, "loss": 0.0022, "step": 173000 }, { "epoch": 17.41, "eval_cer": 0.05777121899887297, "eval_loss": 0.5593122243881226, "eval_runtime": 1631.0957, "eval_samples_per_second": 5.0, "eval_steps_per_second": 0.625, "step": 173000 }, { "epoch": 17.41, "learning_rate": 6.499293856552002e-06, "loss": 0.0021, "step": 173050 }, { "epoch": 17.42, "learning_rate": 6.486684152123475e-06, "loss": 0.0021, "step": 173100 }, { "epoch": 17.42, "learning_rate": 6.474074447694946e-06, "loss": 0.0016, "step": 173150 }, { "epoch": 17.43, "learning_rate": 6.4614647432664185e-06, "loss": 0.0018, "step": 173200 }, { "epoch": 17.43, "learning_rate": 6.448855038837889e-06, "loss": 0.002, "step": 173250 }, { "epoch": 17.44, "learning_rate": 6.4362453344093615e-06, "loss": 0.0018, "step": 173300 }, { "epoch": 17.44, "learning_rate": 6.423635629980834e-06, "loss": 0.0015, "step": 173350 }, { "epoch": 17.45, "learning_rate": 6.411025925552305e-06, "loss": 0.0014, "step": 173400 }, { "epoch": 17.45, "learning_rate": 6.398416221123778e-06, "loss": 0.0032, "step": 173450 }, { "epoch": 17.46, "learning_rate": 6.38605871078382e-06, "loss": 0.002, "step": 173500 }, { "epoch": 17.46, "eval_cer": 0.05777433234328981, "eval_loss": 0.5575246214866638, "eval_runtime": 1333.8273, "eval_samples_per_second": 6.114, "eval_steps_per_second": 0.765, "step": 173500 }, { "epoch": 17.46, "learning_rate": 6.3734490063552904e-06, "loss": 0.0028, "step": 173550 }, { "epoch": 17.47, "learning_rate": 6.360839301926763e-06, "loss": 0.003, "step": 173600 }, { "epoch": 17.47, "learning_rate": 6.348229597498236e-06, "loss": 0.0021, "step": 173650 }, { "epoch": 17.48, "learning_rate": 6.335619893069707e-06, "loss": 0.0022, "step": 173700 }, { "epoch": 17.48, "learning_rate": 6.323010188641179e-06, "loss": 0.0017, "step": 173750 }, { "epoch": 17.49, "learning_rate": 6.31040048421265e-06, "loss": 0.0012, "step": 173800 }, { "epoch": 17.49, "learning_rate": 6.297790779784123e-06, "loss": 0.0025, "step": 173850 }, { "epoch": 17.5, "learning_rate": 6.285181075355593e-06, "loss": 0.0016, "step": 173900 }, { "epoch": 17.5, "learning_rate": 6.2725713709270665e-06, "loss": 0.0015, "step": 173950 }, { "epoch": 17.51, "learning_rate": 6.259961666498537e-06, "loss": 0.002, "step": 174000 }, { "epoch": 17.51, "eval_cer": 0.05795179297504966, "eval_loss": 0.5577999353408813, "eval_runtime": 1300.9708, "eval_samples_per_second": 6.268, "eval_steps_per_second": 0.784, "step": 174000 }, { "epoch": 17.51, "learning_rate": 6.2473519620700095e-06, "loss": 0.0022, "step": 174050 }, { "epoch": 17.52, "learning_rate": 6.234742257641482e-06, "loss": 0.0015, "step": 174100 }, { "epoch": 17.52, "learning_rate": 6.222132553212953e-06, "loss": 0.0016, "step": 174150 }, { "epoch": 17.53, "learning_rate": 6.209522848784425e-06, "loss": 0.0015, "step": 174200 }, { "epoch": 17.53, "learning_rate": 6.196913144355896e-06, "loss": 0.0027, "step": 174250 }, { "epoch": 17.54, "learning_rate": 6.184555634015939e-06, "loss": 0.0014, "step": 174300 }, { "epoch": 17.54, "learning_rate": 6.171945929587411e-06, "loss": 0.0016, "step": 174350 }, { "epoch": 17.55, "learning_rate": 6.159336225158883e-06, "loss": 0.0024, "step": 174400 }, { "epoch": 17.55, "learning_rate": 6.146726520730355e-06, "loss": 0.0016, "step": 174450 }, { "epoch": 17.56, "learning_rate": 6.134116816301826e-06, "loss": 0.0021, "step": 174500 }, { "epoch": 17.56, "eval_cer": 0.0578210325095424, "eval_loss": 0.5583814978599548, "eval_runtime": 1307.0953, "eval_samples_per_second": 6.239, "eval_steps_per_second": 0.78, "step": 174500 }, { "epoch": 17.56, "learning_rate": 6.121507111873298e-06, "loss": 0.0018, "step": 174550 }, { "epoch": 17.57, "learning_rate": 6.10889740744477e-06, "loss": 0.0019, "step": 174600 }, { "epoch": 17.57, "learning_rate": 6.096287703016241e-06, "loss": 0.0021, "step": 174650 }, { "epoch": 17.58, "learning_rate": 6.083677998587714e-06, "loss": 0.0018, "step": 174700 }, { "epoch": 17.58, "learning_rate": 6.071068294159185e-06, "loss": 0.0016, "step": 174750 }, { "epoch": 17.59, "learning_rate": 6.058458589730657e-06, "loss": 0.0019, "step": 174800 }, { "epoch": 17.59, "learning_rate": 6.045848885302129e-06, "loss": 0.0018, "step": 174850 }, { "epoch": 17.6, "learning_rate": 6.0332391808736005e-06, "loss": 0.002, "step": 174900 }, { "epoch": 17.6, "learning_rate": 6.020629476445072e-06, "loss": 0.0028, "step": 174950 }, { "epoch": 17.61, "learning_rate": 6.008019772016544e-06, "loss": 0.0016, "step": 175000 }, { "epoch": 17.61, "eval_cer": 0.05764045853336571, "eval_loss": 0.5597366690635681, "eval_runtime": 1411.5836, "eval_samples_per_second": 5.777, "eval_steps_per_second": 0.723, "step": 175000 }, { "epoch": 17.61, "learning_rate": 5.995410067588017e-06, "loss": 0.0027, "step": 175050 }, { "epoch": 17.62, "learning_rate": 5.982800363159488e-06, "loss": 0.0021, "step": 175100 }, { "epoch": 17.62, "learning_rate": 5.97019065873096e-06, "loss": 0.0018, "step": 175150 }, { "epoch": 17.63, "learning_rate": 5.957580954302432e-06, "loss": 0.0019, "step": 175200 }, { "epoch": 17.63, "learning_rate": 5.944971249873903e-06, "loss": 0.0019, "step": 175250 }, { "epoch": 17.64, "learning_rate": 5.932361545445375e-06, "loss": 0.0019, "step": 175300 }, { "epoch": 17.64, "learning_rate": 5.919751841016847e-06, "loss": 0.0024, "step": 175350 }, { "epoch": 17.65, "learning_rate": 5.907142136588319e-06, "loss": 0.0022, "step": 175400 }, { "epoch": 17.65, "learning_rate": 5.89453243215979e-06, "loss": 0.0023, "step": 175450 }, { "epoch": 17.66, "learning_rate": 5.8819227277312625e-06, "loss": 0.0018, "step": 175500 }, { "epoch": 17.66, "eval_cer": 0.05785839264254448, "eval_loss": 0.5598279237747192, "eval_runtime": 1636.6826, "eval_samples_per_second": 4.983, "eval_steps_per_second": 0.623, "step": 175500 }, { "epoch": 17.66, "learning_rate": 5.869313023302734e-06, "loss": 0.0015, "step": 175550 }, { "epoch": 17.67, "learning_rate": 5.8567033188742055e-06, "loss": 0.0015, "step": 175600 }, { "epoch": 17.67, "learning_rate": 5.844093614445678e-06, "loss": 0.0023, "step": 175650 }, { "epoch": 17.68, "learning_rate": 5.831483910017149e-06, "loss": 0.0013, "step": 175700 }, { "epoch": 17.68, "learning_rate": 5.818874205588621e-06, "loss": 0.002, "step": 175750 }, { "epoch": 17.69, "learning_rate": 5.806264501160093e-06, "loss": 0.0014, "step": 175800 }, { "epoch": 17.69, "learning_rate": 5.7936547967315655e-06, "loss": 0.0027, "step": 175850 }, { "epoch": 17.7, "learning_rate": 5.781045092303037e-06, "loss": 0.0018, "step": 175900 }, { "epoch": 17.7, "learning_rate": 5.768435387874508e-06, "loss": 0.0023, "step": 175950 }, { "epoch": 17.71, "learning_rate": 5.755825683445981e-06, "loss": 0.0028, "step": 176000 }, { "epoch": 17.71, "eval_cer": 0.05772763217703722, "eval_loss": 0.5597357749938965, "eval_runtime": 1665.4346, "eval_samples_per_second": 4.897, "eval_steps_per_second": 0.612, "step": 176000 }, { "epoch": 17.71, "learning_rate": 5.743215979017452e-06, "loss": 0.002, "step": 176050 }, { "epoch": 17.72, "learning_rate": 5.730606274588924e-06, "loss": 0.0015, "step": 176100 }, { "epoch": 17.72, "learning_rate": 5.717996570160396e-06, "loss": 0.0027, "step": 176150 }, { "epoch": 17.73, "learning_rate": 5.7053868657318675e-06, "loss": 0.0017, "step": 176200 }, { "epoch": 17.73, "learning_rate": 5.69302935539191e-06, "loss": 0.0023, "step": 176250 }, { "epoch": 17.74, "learning_rate": 5.680419650963382e-06, "loss": 0.0019, "step": 176300 }, { "epoch": 17.74, "learning_rate": 5.6678099465348535e-06, "loss": 0.0022, "step": 176350 }, { "epoch": 17.75, "learning_rate": 5.655200242106325e-06, "loss": 0.0017, "step": 176400 }, { "epoch": 17.76, "learning_rate": 5.642590537677797e-06, "loss": 0.0022, "step": 176450 }, { "epoch": 17.76, "learning_rate": 5.629980833249269e-06, "loss": 0.002, "step": 176500 }, { "epoch": 17.76, "eval_cer": 0.05745988455718903, "eval_loss": 0.5584138631820679, "eval_runtime": 1639.098, "eval_samples_per_second": 4.975, "eval_steps_per_second": 0.622, "step": 176500 }, { "epoch": 17.77, "learning_rate": 5.61737112882074e-06, "loss": 0.0018, "step": 176550 }, { "epoch": 17.77, "learning_rate": 5.604761424392213e-06, "loss": 0.0016, "step": 176600 }, { "epoch": 17.78, "learning_rate": 5.592151719963684e-06, "loss": 0.0021, "step": 176650 }, { "epoch": 17.78, "learning_rate": 5.579542015535156e-06, "loss": 0.0015, "step": 176700 }, { "epoch": 17.79, "learning_rate": 5.566932311106628e-06, "loss": 0.002, "step": 176750 }, { "epoch": 17.79, "learning_rate": 5.554322606678099e-06, "loss": 0.0029, "step": 176800 }, { "epoch": 17.8, "learning_rate": 5.541712902249572e-06, "loss": 0.0019, "step": 176850 }, { "epoch": 17.8, "learning_rate": 5.529103197821044e-06, "loss": 0.0017, "step": 176900 }, { "epoch": 17.81, "learning_rate": 5.5164934933925155e-06, "loss": 0.0017, "step": 176950 }, { "epoch": 17.81, "learning_rate": 5.503883788963987e-06, "loss": 0.0021, "step": 177000 }, { "epoch": 17.81, "eval_cer": 0.057752538932371936, "eval_loss": 0.5598008632659912, "eval_runtime": 1475.815, "eval_samples_per_second": 5.526, "eval_steps_per_second": 0.691, "step": 177000 }, { "epoch": 17.82, "learning_rate": 5.4912740845354585e-06, "loss": 0.0015, "step": 177050 }, { "epoch": 17.82, "learning_rate": 5.478664380106931e-06, "loss": 0.0017, "step": 177100 }, { "epoch": 17.83, "learning_rate": 5.466054675678402e-06, "loss": 0.0029, "step": 177150 }, { "epoch": 17.83, "learning_rate": 5.453444971249874e-06, "loss": 0.0018, "step": 177200 }, { "epoch": 17.84, "learning_rate": 5.440835266821346e-06, "loss": 0.0024, "step": 177250 }, { "epoch": 17.84, "learning_rate": 5.428225562392818e-06, "loss": 0.0012, "step": 177300 }, { "epoch": 17.85, "learning_rate": 5.415615857964289e-06, "loss": 0.0019, "step": 177350 }, { "epoch": 17.85, "learning_rate": 5.4030061535357614e-06, "loss": 0.0016, "step": 177400 }, { "epoch": 17.86, "learning_rate": 5.390396449107233e-06, "loss": 0.0022, "step": 177450 }, { "epoch": 17.86, "learning_rate": 5.377786744678704e-06, "loss": 0.0019, "step": 177500 }, { "epoch": 17.86, "eval_cer": 0.05713921008225456, "eval_loss": 0.5596818923950195, "eval_runtime": 1311.4477, "eval_samples_per_second": 6.218, "eval_steps_per_second": 0.778, "step": 177500 }, { "epoch": 17.87, "learning_rate": 5.365177040250177e-06, "loss": 0.0016, "step": 177550 }, { "epoch": 17.87, "learning_rate": 5.352567335821648e-06, "loss": 0.0018, "step": 177600 }, { "epoch": 17.88, "learning_rate": 5.3399576313931205e-06, "loss": 0.0016, "step": 177650 }, { "epoch": 17.88, "learning_rate": 5.327347926964593e-06, "loss": 0.0024, "step": 177700 }, { "epoch": 17.89, "learning_rate": 5.314738222536064e-06, "loss": 0.002, "step": 177750 }, { "epoch": 17.89, "learning_rate": 5.302128518107536e-06, "loss": 0.003, "step": 177800 }, { "epoch": 17.9, "learning_rate": 5.289518813679008e-06, "loss": 0.0019, "step": 177850 }, { "epoch": 17.9, "learning_rate": 5.27690910925048e-06, "loss": 0.0017, "step": 177900 }, { "epoch": 17.91, "learning_rate": 5.264299404821951e-06, "loss": 0.0014, "step": 177950 }, { "epoch": 17.91, "learning_rate": 5.251689700393423e-06, "loss": 0.0015, "step": 178000 }, { "epoch": 17.91, "eval_cer": 0.05736025753585016, "eval_loss": 0.558789849281311, "eval_runtime": 1310.7157, "eval_samples_per_second": 6.222, "eval_steps_per_second": 0.778, "step": 178000 }, { "epoch": 17.92, "learning_rate": 5.239079995964895e-06, "loss": 0.002, "step": 178050 }, { "epoch": 17.92, "learning_rate": 5.2264702915363664e-06, "loss": 0.0021, "step": 178100 }, { "epoch": 17.93, "learning_rate": 5.213860587107838e-06, "loss": 0.0025, "step": 178150 }, { "epoch": 17.93, "learning_rate": 5.20125088267931e-06, "loss": 0.0015, "step": 178200 }, { "epoch": 17.94, "learning_rate": 5.188641178250782e-06, "loss": 0.0018, "step": 178250 }, { "epoch": 17.94, "learning_rate": 5.176031473822253e-06, "loss": 0.0016, "step": 178300 }, { "epoch": 17.95, "learning_rate": 5.1634217693937255e-06, "loss": 0.0018, "step": 178350 }, { "epoch": 17.95, "learning_rate": 5.150812064965198e-06, "loss": 0.0018, "step": 178400 }, { "epoch": 17.96, "learning_rate": 5.138202360536669e-06, "loss": 0.0015, "step": 178450 }, { "epoch": 17.96, "learning_rate": 5.125592656108142e-06, "loss": 0.0025, "step": 178500 }, { "epoch": 17.96, "eval_cer": 0.057335350780515444, "eval_loss": 0.5605639815330505, "eval_runtime": 1314.8291, "eval_samples_per_second": 6.202, "eval_steps_per_second": 0.776, "step": 178500 }, { "epoch": 17.97, "learning_rate": 5.112982951679613e-06, "loss": 0.0012, "step": 178550 }, { "epoch": 17.97, "learning_rate": 5.100373247251085e-06, "loss": 0.0028, "step": 178600 }, { "epoch": 17.98, "learning_rate": 5.087763542822557e-06, "loss": 0.0016, "step": 178650 }, { "epoch": 17.98, "learning_rate": 5.0751538383940285e-06, "loss": 0.0013, "step": 178700 }, { "epoch": 17.99, "learning_rate": 5.0625441339655e-06, "loss": 0.0014, "step": 178750 }, { "epoch": 17.99, "learning_rate": 5.049934429536972e-06, "loss": 0.0022, "step": 178800 }, { "epoch": 18.0, "learning_rate": 5.037324725108444e-06, "loss": 0.0019, "step": 178850 }, { "epoch": 18.0, "learning_rate": 5.024715020679915e-06, "loss": 0.0025, "step": 178900 }, { "epoch": 18.01, "learning_rate": 5.012105316251387e-06, "loss": 0.0011, "step": 178950 }, { "epoch": 18.01, "learning_rate": 4.999495611822859e-06, "loss": 0.0013, "step": 179000 }, { "epoch": 18.01, "eval_cer": 0.05693684269515999, "eval_loss": 0.5582064390182495, "eval_runtime": 1312.3741, "eval_samples_per_second": 6.214, "eval_steps_per_second": 0.777, "step": 179000 }, { "epoch": 18.02, "learning_rate": 4.9868859073943305e-06, "loss": 0.0012, "step": 179050 }, { "epoch": 18.02, "learning_rate": 4.974276202965802e-06, "loss": 0.0012, "step": 179100 }, { "epoch": 18.03, "learning_rate": 4.961666498537274e-06, "loss": 0.0019, "step": 179150 }, { "epoch": 18.03, "learning_rate": 4.949056794108747e-06, "loss": 0.0013, "step": 179200 }, { "epoch": 18.04, "learning_rate": 4.936447089680218e-06, "loss": 0.0016, "step": 179250 }, { "epoch": 18.04, "learning_rate": 4.9238373852516905e-06, "loss": 0.0015, "step": 179300 }, { "epoch": 18.05, "learning_rate": 4.911227680823162e-06, "loss": 0.0014, "step": 179350 }, { "epoch": 18.05, "learning_rate": 4.8986179763946335e-06, "loss": 0.0017, "step": 179400 }, { "epoch": 18.06, "learning_rate": 4.886008271966106e-06, "loss": 0.001, "step": 179450 }, { "epoch": 18.06, "learning_rate": 4.873398567537577e-06, "loss": 0.001, "step": 179500 }, { "epoch": 18.06, "eval_cer": 0.05759064502269628, "eval_loss": 0.5600205659866333, "eval_runtime": 1343.0882, "eval_samples_per_second": 6.072, "eval_steps_per_second": 0.759, "step": 179500 }, { "epoch": 18.07, "learning_rate": 4.860788863109049e-06, "loss": 0.0012, "step": 179550 }, { "epoch": 18.07, "learning_rate": 4.848179158680521e-06, "loss": 0.0024, "step": 179600 }, { "epoch": 18.08, "learning_rate": 4.835569454251993e-06, "loss": 0.0011, "step": 179650 }, { "epoch": 18.08, "learning_rate": 4.822959749823464e-06, "loss": 0.0013, "step": 179700 }, { "epoch": 18.09, "learning_rate": 4.810350045394936e-06, "loss": 0.0011, "step": 179750 }, { "epoch": 18.09, "learning_rate": 4.797740340966408e-06, "loss": 0.0016, "step": 179800 }, { "epoch": 18.1, "learning_rate": 4.785130636537879e-06, "loss": 0.0011, "step": 179850 }, { "epoch": 18.1, "learning_rate": 4.772520932109352e-06, "loss": 0.0012, "step": 179900 }, { "epoch": 18.11, "learning_rate": 4.759911227680824e-06, "loss": 0.0012, "step": 179950 }, { "epoch": 18.11, "learning_rate": 4.7473015232522955e-06, "loss": 0.0015, "step": 180000 }, { "epoch": 18.11, "eval_cer": 0.05737271091351752, "eval_loss": 0.558866560459137, "eval_runtime": 1312.7141, "eval_samples_per_second": 6.212, "eval_steps_per_second": 0.777, "step": 180000 }, { "epoch": 18.12, "learning_rate": 4.734691818823767e-06, "loss": 0.0015, "step": 180050 }, { "epoch": 18.12, "learning_rate": 4.722082114395239e-06, "loss": 0.0013, "step": 180100 }, { "epoch": 18.13, "learning_rate": 4.709472409966711e-06, "loss": 0.0011, "step": 180150 }, { "epoch": 18.13, "learning_rate": 4.696862705538182e-06, "loss": 0.0009, "step": 180200 }, { "epoch": 18.14, "learning_rate": 4.684253001109655e-06, "loss": 0.0012, "step": 180250 }, { "epoch": 18.14, "learning_rate": 4.671643296681126e-06, "loss": 0.002, "step": 180300 }, { "epoch": 18.15, "learning_rate": 4.659033592252598e-06, "loss": 0.0015, "step": 180350 }, { "epoch": 18.15, "learning_rate": 4.64642388782407e-06, "loss": 0.0012, "step": 180400 }, { "epoch": 18.16, "learning_rate": 4.633814183395541e-06, "loss": 0.0009, "step": 180450 }, { "epoch": 18.16, "learning_rate": 4.621204478967013e-06, "loss": 0.0008, "step": 180500 }, { "epoch": 18.16, "eval_cer": 0.057263743858928136, "eval_loss": 0.561207115650177, "eval_runtime": 1324.1115, "eval_samples_per_second": 6.159, "eval_steps_per_second": 0.77, "step": 180500 }, { "epoch": 18.17, "learning_rate": 4.608594774538485e-06, "loss": 0.0013, "step": 180550 }, { "epoch": 18.17, "learning_rate": 4.595985070109957e-06, "loss": 0.001, "step": 180600 }, { "epoch": 18.18, "learning_rate": 4.583375365681428e-06, "loss": 0.0013, "step": 180650 }, { "epoch": 18.18, "learning_rate": 4.5707656612529005e-06, "loss": 0.0012, "step": 180700 }, { "epoch": 18.19, "learning_rate": 4.558408150912943e-06, "loss": 0.0022, "step": 180750 }, { "epoch": 18.19, "learning_rate": 4.545798446484414e-06, "loss": 0.0012, "step": 180800 }, { "epoch": 18.2, "learning_rate": 4.5331887420558865e-06, "loss": 0.0016, "step": 180850 }, { "epoch": 18.2, "learning_rate": 4.520579037627358e-06, "loss": 0.001, "step": 180900 }, { "epoch": 18.21, "learning_rate": 4.5079693331988295e-06, "loss": 0.0009, "step": 180950 }, { "epoch": 18.21, "learning_rate": 4.495359628770302e-06, "loss": 0.0009, "step": 181000 }, { "epoch": 18.21, "eval_cer": 0.05701156296116414, "eval_loss": 0.5615878105163574, "eval_runtime": 1309.1728, "eval_samples_per_second": 6.229, "eval_steps_per_second": 0.779, "step": 181000 }, { "epoch": 18.22, "learning_rate": 4.482749924341774e-06, "loss": 0.0011, "step": 181050 }, { "epoch": 18.22, "learning_rate": 4.470140219913246e-06, "loss": 0.0013, "step": 181100 }, { "epoch": 18.23, "learning_rate": 4.457530515484717e-06, "loss": 0.0014, "step": 181150 }, { "epoch": 18.23, "learning_rate": 4.444920811056189e-06, "loss": 0.0011, "step": 181200 }, { "epoch": 18.24, "learning_rate": 4.432311106627661e-06, "loss": 0.0012, "step": 181250 }, { "epoch": 18.24, "learning_rate": 4.419701402199132e-06, "loss": 0.0013, "step": 181300 }, { "epoch": 18.25, "learning_rate": 4.407091697770605e-06, "loss": 0.0014, "step": 181350 }, { "epoch": 18.25, "learning_rate": 4.394481993342076e-06, "loss": 0.0011, "step": 181400 }, { "epoch": 18.26, "learning_rate": 4.381872288913548e-06, "loss": 0.0008, "step": 181450 }, { "epoch": 18.26, "learning_rate": 4.36926258448502e-06, "loss": 0.0013, "step": 181500 }, { "epoch": 18.26, "eval_cer": 0.05707694319391777, "eval_loss": 0.5610823035240173, "eval_runtime": 1351.8632, "eval_samples_per_second": 6.032, "eval_steps_per_second": 0.755, "step": 181500 }, { "epoch": 18.27, "learning_rate": 4.3566528800564915e-06, "loss": 0.0013, "step": 181550 }, { "epoch": 18.27, "learning_rate": 4.344043175627963e-06, "loss": 0.0014, "step": 181600 }, { "epoch": 18.28, "learning_rate": 4.331433471199435e-06, "loss": 0.0008, "step": 181650 }, { "epoch": 18.28, "learning_rate": 4.318823766770907e-06, "loss": 0.0018, "step": 181700 }, { "epoch": 18.29, "learning_rate": 4.306214062342378e-06, "loss": 0.001, "step": 181750 }, { "epoch": 18.29, "learning_rate": 4.293604357913851e-06, "loss": 0.0016, "step": 181800 }, { "epoch": 18.3, "learning_rate": 4.280994653485323e-06, "loss": 0.0009, "step": 181850 }, { "epoch": 18.3, "learning_rate": 4.268384949056794e-06, "loss": 0.0008, "step": 181900 }, { "epoch": 18.31, "learning_rate": 4.255775244628266e-06, "loss": 0.0009, "step": 181950 }, { "epoch": 18.31, "learning_rate": 4.243165540199738e-06, "loss": 0.0016, "step": 182000 }, { "epoch": 18.31, "eval_cer": 0.05728865061426285, "eval_loss": 0.5610185265541077, "eval_runtime": 1325.6974, "eval_samples_per_second": 6.151, "eval_steps_per_second": 0.769, "step": 182000 }, { "epoch": 18.32, "learning_rate": 4.23055583577121e-06, "loss": 0.0014, "step": 182050 }, { "epoch": 18.32, "learning_rate": 4.217946131342681e-06, "loss": 0.0014, "step": 182100 }, { "epoch": 18.33, "learning_rate": 4.2053364269141535e-06, "loss": 0.0009, "step": 182150 }, { "epoch": 18.33, "learning_rate": 4.192726722485625e-06, "loss": 0.0014, "step": 182200 }, { "epoch": 18.34, "learning_rate": 4.1801170180570965e-06, "loss": 0.0017, "step": 182250 }, { "epoch": 18.34, "learning_rate": 4.167507313628569e-06, "loss": 0.002, "step": 182300 }, { "epoch": 18.35, "learning_rate": 4.15489760920004e-06, "loss": 0.001, "step": 182350 }, { "epoch": 18.35, "learning_rate": 4.142287904771512e-06, "loss": 0.001, "step": 182400 }, { "epoch": 18.36, "learning_rate": 4.129678200342984e-06, "loss": 0.002, "step": 182450 }, { "epoch": 18.36, "learning_rate": 4.117068495914456e-06, "loss": 0.0011, "step": 182500 }, { "epoch": 18.36, "eval_cer": 0.05703024302766518, "eval_loss": 0.5605985522270203, "eval_runtime": 1310.3686, "eval_samples_per_second": 6.223, "eval_steps_per_second": 0.778, "step": 182500 }, { "epoch": 18.37, "learning_rate": 4.104458791485928e-06, "loss": 0.0011, "step": 182550 }, { "epoch": 18.37, "learning_rate": 4.0918490870574e-06, "loss": 0.0008, "step": 182600 }, { "epoch": 18.38, "learning_rate": 4.079239382628872e-06, "loss": 0.0012, "step": 182650 }, { "epoch": 18.38, "learning_rate": 4.066629678200343e-06, "loss": 0.0008, "step": 182700 }, { "epoch": 18.39, "learning_rate": 4.054019973771815e-06, "loss": 0.0015, "step": 182750 }, { "epoch": 18.39, "learning_rate": 4.041410269343287e-06, "loss": 0.0013, "step": 182800 }, { "epoch": 18.4, "learning_rate": 4.0288005649147585e-06, "loss": 0.0011, "step": 182850 }, { "epoch": 18.4, "learning_rate": 4.01619086048623e-06, "loss": 0.0016, "step": 182900 }, { "epoch": 18.41, "learning_rate": 4.003581156057702e-06, "loss": 0.0078, "step": 182950 }, { "epoch": 18.41, "learning_rate": 3.990971451629174e-06, "loss": 0.0014, "step": 183000 }, { "epoch": 18.41, "eval_cer": 0.05716723018200612, "eval_loss": 0.5614513754844666, "eval_runtime": 1311.5515, "eval_samples_per_second": 6.218, "eval_steps_per_second": 0.778, "step": 183000 }, { "epoch": 18.42, "learning_rate": 3.978361747200645e-06, "loss": 0.0012, "step": 183050 }, { "epoch": 18.42, "learning_rate": 3.965752042772118e-06, "loss": 0.0013, "step": 183100 }, { "epoch": 18.43, "learning_rate": 3.953142338343589e-06, "loss": 0.0009, "step": 183150 }, { "epoch": 18.43, "learning_rate": 3.940532633915061e-06, "loss": 0.0013, "step": 183200 }, { "epoch": 18.44, "learning_rate": 3.927922929486533e-06, "loss": 0.0009, "step": 183250 }, { "epoch": 18.44, "learning_rate": 3.915313225058004e-06, "loss": 0.001, "step": 183300 }, { "epoch": 18.45, "learning_rate": 3.902703520629477e-06, "loss": 0.0015, "step": 183350 }, { "epoch": 18.45, "learning_rate": 3.890093816200949e-06, "loss": 0.0015, "step": 183400 }, { "epoch": 18.46, "learning_rate": 3.8774841117724206e-06, "loss": 0.0011, "step": 183450 }, { "epoch": 18.46, "learning_rate": 3.864874407343892e-06, "loss": 0.0008, "step": 183500 }, { "epoch": 18.46, "eval_cer": 0.05679362885198538, "eval_loss": 0.5621600151062012, "eval_runtime": 1308.627, "eval_samples_per_second": 6.232, "eval_steps_per_second": 0.779, "step": 183500 }, { "epoch": 18.47, "learning_rate": 3.852264702915364e-06, "loss": 0.0011, "step": 183550 }, { "epoch": 18.47, "learning_rate": 3.839654998486836e-06, "loss": 0.0016, "step": 183600 }, { "epoch": 18.48, "learning_rate": 3.827045294058307e-06, "loss": 0.001, "step": 183650 }, { "epoch": 18.48, "learning_rate": 3.8144355896297793e-06, "loss": 0.001, "step": 183700 }, { "epoch": 18.49, "learning_rate": 3.802078079289822e-06, "loss": 0.0017, "step": 183750 }, { "epoch": 18.49, "learning_rate": 3.7894683748612937e-06, "loss": 0.0007, "step": 183800 }, { "epoch": 18.5, "learning_rate": 3.7768586704327652e-06, "loss": 0.0009, "step": 183850 }, { "epoch": 18.5, "learning_rate": 3.764248966004237e-06, "loss": 0.0012, "step": 183900 }, { "epoch": 18.51, "learning_rate": 3.751639261575709e-06, "loss": 0.0012, "step": 183950 }, { "epoch": 18.51, "learning_rate": 3.7390295571471805e-06, "loss": 0.0011, "step": 184000 }, { "epoch": 18.51, "eval_cer": 0.05688702918449056, "eval_loss": 0.562957763671875, "eval_runtime": 1309.7525, "eval_samples_per_second": 6.226, "eval_steps_per_second": 0.779, "step": 184000 }, { "epoch": 18.52, "learning_rate": 3.7264198527186524e-06, "loss": 0.0012, "step": 184050 }, { "epoch": 18.52, "learning_rate": 3.7138101482901243e-06, "loss": 0.0013, "step": 184100 }, { "epoch": 18.53, "learning_rate": 3.701200443861596e-06, "loss": 0.0023, "step": 184150 }, { "epoch": 18.53, "learning_rate": 3.6885907394330677e-06, "loss": 0.0019, "step": 184200 }, { "epoch": 18.54, "learning_rate": 3.675981035004539e-06, "loss": 0.0015, "step": 184250 }, { "epoch": 18.54, "learning_rate": 3.663371330576011e-06, "loss": 0.0015, "step": 184300 }, { "epoch": 18.55, "learning_rate": 3.650761626147483e-06, "loss": 0.001, "step": 184350 }, { "epoch": 18.56, "learning_rate": 3.6381519217189554e-06, "loss": 0.0012, "step": 184400 }, { "epoch": 18.56, "learning_rate": 3.6255422172904273e-06, "loss": 0.0011, "step": 184450 }, { "epoch": 18.57, "learning_rate": 3.6129325128618987e-06, "loss": 0.0011, "step": 184500 }, { "epoch": 18.57, "eval_cer": 0.057145436771088236, "eval_loss": 0.5615208745002747, "eval_runtime": 1349.1712, "eval_samples_per_second": 6.044, "eval_steps_per_second": 0.756, "step": 184500 }, { "epoch": 18.57, "learning_rate": 3.6003228084333707e-06, "loss": 0.0014, "step": 184550 }, { "epoch": 18.58, "learning_rate": 3.5877131040048426e-06, "loss": 0.0014, "step": 184600 }, { "epoch": 18.58, "learning_rate": 3.575103399576314e-06, "loss": 0.0012, "step": 184650 }, { "epoch": 18.59, "learning_rate": 3.562493695147786e-06, "loss": 0.0013, "step": 184700 }, { "epoch": 18.59, "learning_rate": 3.549883990719258e-06, "loss": 0.0012, "step": 184750 }, { "epoch": 18.6, "learning_rate": 3.5372742862907293e-06, "loss": 0.0014, "step": 184800 }, { "epoch": 18.6, "learning_rate": 3.5246645818622012e-06, "loss": 0.0017, "step": 184850 }, { "epoch": 18.61, "learning_rate": 3.512054877433673e-06, "loss": 0.0013, "step": 184900 }, { "epoch": 18.61, "learning_rate": 3.4994451730051446e-06, "loss": 0.0019, "step": 184950 }, { "epoch": 18.62, "learning_rate": 3.4868354685766165e-06, "loss": 0.0013, "step": 185000 }, { "epoch": 18.62, "eval_cer": 0.05750658472344162, "eval_loss": 0.5602685809135437, "eval_runtime": 1343.1528, "eval_samples_per_second": 6.072, "eval_steps_per_second": 0.759, "step": 185000 }, { "epoch": 18.62, "learning_rate": 3.4742257641480885e-06, "loss": 0.0014, "step": 185050 }, { "epoch": 18.63, "learning_rate": 3.46161605971956e-06, "loss": 0.0019, "step": 185100 }, { "epoch": 18.63, "learning_rate": 3.4490063552910323e-06, "loss": 0.0012, "step": 185150 }, { "epoch": 18.64, "learning_rate": 3.436396650862504e-06, "loss": 0.0007, "step": 185200 }, { "epoch": 18.64, "learning_rate": 3.423786946433976e-06, "loss": 0.0015, "step": 185250 }, { "epoch": 18.65, "learning_rate": 3.4111772420054476e-06, "loss": 0.0012, "step": 185300 }, { "epoch": 18.65, "learning_rate": 3.3985675375769195e-06, "loss": 0.0011, "step": 185350 }, { "epoch": 18.66, "learning_rate": 3.3859578331483914e-06, "loss": 0.0016, "step": 185400 }, { "epoch": 18.66, "learning_rate": 3.373348128719863e-06, "loss": 0.0021, "step": 185450 }, { "epoch": 18.67, "learning_rate": 3.3607384242913348e-06, "loss": 0.001, "step": 185500 }, { "epoch": 18.67, "eval_cer": 0.0573478041581828, "eval_loss": 0.5609522461891174, "eval_runtime": 1316.5752, "eval_samples_per_second": 6.194, "eval_steps_per_second": 0.775, "step": 185500 }, { "epoch": 18.67, "learning_rate": 3.3481287198628067e-06, "loss": 0.0011, "step": 185550 }, { "epoch": 18.68, "learning_rate": 3.335519015434278e-06, "loss": 0.0016, "step": 185600 }, { "epoch": 18.68, "learning_rate": 3.32290931100575e-06, "loss": 0.0013, "step": 185650 }, { "epoch": 18.69, "learning_rate": 3.310299606577222e-06, "loss": 0.001, "step": 185700 }, { "epoch": 18.69, "learning_rate": 3.2976899021486935e-06, "loss": 0.0019, "step": 185750 }, { "epoch": 18.7, "learning_rate": 3.2850801977201654e-06, "loss": 0.0009, "step": 185800 }, { "epoch": 18.7, "learning_rate": 3.2724704932916373e-06, "loss": 0.0022, "step": 185850 }, { "epoch": 18.71, "learning_rate": 3.2598607888631088e-06, "loss": 0.0014, "step": 185900 }, { "epoch": 18.71, "learning_rate": 3.2472510844345815e-06, "loss": 0.0013, "step": 185950 }, { "epoch": 18.72, "learning_rate": 3.234641380006053e-06, "loss": 0.0015, "step": 186000 }, { "epoch": 18.72, "eval_cer": 0.05756573826736156, "eval_loss": 0.5612325668334961, "eval_runtime": 1408.2915, "eval_samples_per_second": 5.791, "eval_steps_per_second": 0.724, "step": 186000 }, { "epoch": 18.72, "learning_rate": 3.222031675577525e-06, "loss": 0.0009, "step": 186050 }, { "epoch": 18.73, "learning_rate": 3.2094219711489964e-06, "loss": 0.0011, "step": 186100 }, { "epoch": 18.73, "learning_rate": 3.1968122667204683e-06, "loss": 0.0015, "step": 186150 }, { "epoch": 18.74, "learning_rate": 3.18420256229194e-06, "loss": 0.0018, "step": 186200 }, { "epoch": 18.74, "learning_rate": 3.1715928578634117e-06, "loss": 0.001, "step": 186250 }, { "epoch": 18.75, "learning_rate": 3.1589831534348836e-06, "loss": 0.0012, "step": 186300 }, { "epoch": 18.75, "learning_rate": 3.1463734490063555e-06, "loss": 0.0026, "step": 186350 }, { "epoch": 18.76, "learning_rate": 3.133763744577827e-06, "loss": 0.0013, "step": 186400 }, { "epoch": 18.76, "learning_rate": 3.121154040149299e-06, "loss": 0.001, "step": 186450 }, { "epoch": 18.77, "learning_rate": 3.1085443357207712e-06, "loss": 0.0019, "step": 186500 }, { "epoch": 18.77, "eval_cer": 0.05745988455718903, "eval_loss": 0.5607452988624573, "eval_runtime": 1313.2535, "eval_samples_per_second": 6.21, "eval_steps_per_second": 0.777, "step": 186500 }, { "epoch": 18.77, "learning_rate": 3.0959346312922427e-06, "loss": 0.001, "step": 186550 }, { "epoch": 18.78, "learning_rate": 3.0833249268637146e-06, "loss": 0.0018, "step": 186600 }, { "epoch": 18.78, "learning_rate": 3.070715222435186e-06, "loss": 0.0008, "step": 186650 }, { "epoch": 18.79, "learning_rate": 3.058105518006658e-06, "loss": 0.002, "step": 186700 }, { "epoch": 18.79, "learning_rate": 3.04549581357813e-06, "loss": 0.0021, "step": 186750 }, { "epoch": 18.8, "learning_rate": 3.0328861091496014e-06, "loss": 0.0009, "step": 186800 }, { "epoch": 18.8, "learning_rate": 3.0202764047210733e-06, "loss": 0.0009, "step": 186850 }, { "epoch": 18.81, "learning_rate": 3.0076667002925456e-06, "loss": 0.0015, "step": 186900 }, { "epoch": 18.81, "learning_rate": 2.995056995864017e-06, "loss": 0.0011, "step": 186950 }, { "epoch": 18.82, "learning_rate": 2.982447291435489e-06, "loss": 0.0009, "step": 187000 }, { "epoch": 18.82, "eval_cer": 0.057646685222199394, "eval_loss": 0.5606160759925842, "eval_runtime": 1353.0735, "eval_samples_per_second": 6.027, "eval_steps_per_second": 0.754, "step": 187000 }, { "epoch": 18.82, "learning_rate": 2.969837587006961e-06, "loss": 0.0014, "step": 187050 }, { "epoch": 18.83, "learning_rate": 2.9572278825784324e-06, "loss": 0.0012, "step": 187100 }, { "epoch": 18.83, "learning_rate": 2.9446181781499043e-06, "loss": 0.0014, "step": 187150 }, { "epoch": 18.84, "learning_rate": 2.932008473721376e-06, "loss": 0.0012, "step": 187200 }, { "epoch": 18.84, "learning_rate": 2.9193987692928477e-06, "loss": 0.0013, "step": 187250 }, { "epoch": 18.85, "learning_rate": 2.90678906486432e-06, "loss": 0.0014, "step": 187300 }, { "epoch": 18.85, "learning_rate": 2.8941793604357915e-06, "loss": 0.0011, "step": 187350 }, { "epoch": 18.86, "learning_rate": 2.8815696560072634e-06, "loss": 0.0012, "step": 187400 }, { "epoch": 18.86, "learning_rate": 2.8689599515787353e-06, "loss": 0.001, "step": 187450 }, { "epoch": 18.87, "learning_rate": 2.856350247150207e-06, "loss": 0.0012, "step": 187500 }, { "epoch": 18.87, "eval_cer": 0.05721393034825871, "eval_loss": 0.56095951795578, "eval_runtime": 1573.4662, "eval_samples_per_second": 5.183, "eval_steps_per_second": 0.648, "step": 187500 }, { "epoch": 18.87, "learning_rate": 2.8437405427216787e-06, "loss": 0.0011, "step": 187550 }, { "epoch": 18.88, "learning_rate": 2.83113083829315e-06, "loss": 0.0016, "step": 187600 }, { "epoch": 18.88, "learning_rate": 2.8185211338646225e-06, "loss": 0.0022, "step": 187650 }, { "epoch": 18.89, "learning_rate": 2.8059114294360944e-06, "loss": 0.0013, "step": 187700 }, { "epoch": 18.89, "learning_rate": 2.793301725007566e-06, "loss": 0.0011, "step": 187750 }, { "epoch": 18.9, "learning_rate": 2.780692020579038e-06, "loss": 0.0014, "step": 187800 }, { "epoch": 18.9, "learning_rate": 2.7680823161505097e-06, "loss": 0.0023, "step": 187850 }, { "epoch": 18.91, "learning_rate": 2.7554726117219812e-06, "loss": 0.0014, "step": 187900 }, { "epoch": 18.91, "learning_rate": 2.742862907293453e-06, "loss": 0.0015, "step": 187950 }, { "epoch": 18.92, "learning_rate": 2.730253202864925e-06, "loss": 0.0012, "step": 188000 }, { "epoch": 18.92, "eval_cer": 0.057618665122447836, "eval_loss": 0.561955988407135, "eval_runtime": 1594.0671, "eval_samples_per_second": 5.116, "eval_steps_per_second": 0.64, "step": 188000 }, { "epoch": 18.92, "learning_rate": 2.717643498436397e-06, "loss": 0.0013, "step": 188050 }, { "epoch": 18.93, "learning_rate": 2.705285988096439e-06, "loss": 0.0028, "step": 188100 }, { "epoch": 18.93, "learning_rate": 2.692676283667911e-06, "loss": 0.0016, "step": 188150 }, { "epoch": 18.94, "learning_rate": 2.6800665792393825e-06, "loss": 0.0018, "step": 188200 }, { "epoch": 18.94, "learning_rate": 2.6674568748108544e-06, "loss": 0.0021, "step": 188250 }, { "epoch": 18.95, "learning_rate": 2.6548471703823263e-06, "loss": 0.0013, "step": 188300 }, { "epoch": 18.95, "learning_rate": 2.642237465953798e-06, "loss": 0.0014, "step": 188350 }, { "epoch": 18.96, "learning_rate": 2.62962776152527e-06, "loss": 0.0015, "step": 188400 }, { "epoch": 18.96, "learning_rate": 2.617018057096742e-06, "loss": 0.0019, "step": 188450 }, { "epoch": 18.97, "learning_rate": 2.604660546756784e-06, "loss": 0.0025, "step": 188500 }, { "epoch": 18.97, "eval_cer": 0.057382050946768036, "eval_loss": 0.5618418455123901, "eval_runtime": 1654.59, "eval_samples_per_second": 4.929, "eval_steps_per_second": 0.616, "step": 188500 }, { "epoch": 18.97, "learning_rate": 2.5920508423282557e-06, "loss": 0.0016, "step": 188550 }, { "epoch": 18.98, "learning_rate": 2.579441137899728e-06, "loss": 0.001, "step": 188600 }, { "epoch": 18.98, "learning_rate": 2.5668314334711995e-06, "loss": 0.002, "step": 188650 }, { "epoch": 18.99, "learning_rate": 2.5542217290426714e-06, "loss": 0.0017, "step": 188700 }, { "epoch": 18.99, "learning_rate": 2.5416120246141433e-06, "loss": 0.0012, "step": 188750 }, { "epoch": 19.0, "learning_rate": 2.5290023201856148e-06, "loss": 0.0011, "step": 188800 }, { "epoch": 19.0, "learning_rate": 2.5163926157570867e-06, "loss": 0.0008, "step": 188850 }, { "epoch": 19.01, "learning_rate": 2.5037829113285586e-06, "loss": 0.0008, "step": 188900 }, { "epoch": 19.01, "learning_rate": 2.49117320690003e-06, "loss": 0.0008, "step": 188950 }, { "epoch": 19.02, "learning_rate": 2.4785635024715024e-06, "loss": 0.0008, "step": 189000 }, { "epoch": 19.02, "eval_cer": 0.057319784058431245, "eval_loss": 0.5623030066490173, "eval_runtime": 1310.5585, "eval_samples_per_second": 6.223, "eval_steps_per_second": 0.778, "step": 189000 }, { "epoch": 19.02, "learning_rate": 2.465953798042974e-06, "loss": 0.0013, "step": 189050 }, { "epoch": 19.03, "learning_rate": 2.453344093614446e-06, "loss": 0.0009, "step": 189100 }, { "epoch": 19.03, "learning_rate": 2.4407343891859177e-06, "loss": 0.0011, "step": 189150 }, { "epoch": 19.04, "learning_rate": 2.428124684757389e-06, "loss": 0.0014, "step": 189200 }, { "epoch": 19.04, "learning_rate": 2.415514980328861e-06, "loss": 0.0008, "step": 189250 }, { "epoch": 19.05, "learning_rate": 2.402905275900333e-06, "loss": 0.0007, "step": 189300 }, { "epoch": 19.05, "learning_rate": 2.390295571471805e-06, "loss": 0.0012, "step": 189350 }, { "epoch": 19.06, "learning_rate": 2.377685867043277e-06, "loss": 0.0011, "step": 189400 }, { "epoch": 19.06, "learning_rate": 2.3650761626147483e-06, "loss": 0.0008, "step": 189450 }, { "epoch": 19.07, "learning_rate": 2.35246645818622e-06, "loss": 0.0009, "step": 189500 }, { "epoch": 19.07, "eval_cer": 0.05724817713684394, "eval_loss": 0.562041699886322, "eval_runtime": 1329.9267, "eval_samples_per_second": 6.132, "eval_steps_per_second": 0.767, "step": 189500 }, { "epoch": 19.07, "learning_rate": 2.339856753757692e-06, "loss": 0.0009, "step": 189550 }, { "epoch": 19.08, "learning_rate": 2.3272470493291636e-06, "loss": 0.0016, "step": 189600 }, { "epoch": 19.08, "learning_rate": 2.3146373449006355e-06, "loss": 0.0008, "step": 189650 }, { "epoch": 19.09, "learning_rate": 2.3020276404721074e-06, "loss": 0.0011, "step": 189700 }, { "epoch": 19.09, "learning_rate": 2.2894179360435793e-06, "loss": 0.0008, "step": 189750 }, { "epoch": 19.1, "learning_rate": 2.2768082316150512e-06, "loss": 0.0011, "step": 189800 }, { "epoch": 19.1, "learning_rate": 2.264198527186523e-06, "loss": 0.0009, "step": 189850 }, { "epoch": 19.11, "learning_rate": 2.2515888227579946e-06, "loss": 0.0017, "step": 189900 }, { "epoch": 19.11, "learning_rate": 2.2389791183294665e-06, "loss": 0.0007, "step": 189950 }, { "epoch": 19.12, "learning_rate": 2.226369413900938e-06, "loss": 0.001, "step": 190000 }, { "epoch": 19.12, "eval_cer": 0.05724195044801026, "eval_loss": 0.5626487135887146, "eval_runtime": 1310.4772, "eval_samples_per_second": 6.223, "eval_steps_per_second": 0.778, "step": 190000 }, { "epoch": 19.12, "learning_rate": 2.21375970947241e-06, "loss": 0.0008, "step": 190050 }, { "epoch": 19.13, "learning_rate": 2.201150005043882e-06, "loss": 0.0009, "step": 190100 }, { "epoch": 19.13, "learning_rate": 2.1885403006153537e-06, "loss": 0.0013, "step": 190150 }, { "epoch": 19.14, "learning_rate": 2.1759305961868256e-06, "loss": 0.001, "step": 190200 }, { "epoch": 19.14, "learning_rate": 2.1633208917582975e-06, "loss": 0.0007, "step": 190250 }, { "epoch": 19.15, "learning_rate": 2.150711187329769e-06, "loss": 0.0012, "step": 190300 }, { "epoch": 19.15, "learning_rate": 2.138101482901241e-06, "loss": 0.0016, "step": 190350 }, { "epoch": 19.16, "learning_rate": 2.1254917784727124e-06, "loss": 0.001, "step": 190400 }, { "epoch": 19.16, "learning_rate": 2.1128820740441843e-06, "loss": 0.001, "step": 190450 }, { "epoch": 19.17, "learning_rate": 2.1002723696156562e-06, "loss": 0.0009, "step": 190500 }, { "epoch": 19.17, "eval_cer": 0.05705203643858305, "eval_loss": 0.5623518824577332, "eval_runtime": 1311.3606, "eval_samples_per_second": 6.219, "eval_steps_per_second": 0.778, "step": 190500 }, { "epoch": 19.17, "learning_rate": 2.087662665187128e-06, "loss": 0.0014, "step": 190550 }, { "epoch": 19.18, "learning_rate": 2.0750529607586e-06, "loss": 0.0007, "step": 190600 }, { "epoch": 19.18, "learning_rate": 2.062443256330072e-06, "loss": 0.0011, "step": 190650 }, { "epoch": 19.19, "learning_rate": 2.0498335519015434e-06, "loss": 0.001, "step": 190700 }, { "epoch": 19.19, "learning_rate": 2.0372238474730153e-06, "loss": 0.0013, "step": 190750 }, { "epoch": 19.2, "learning_rate": 2.0246141430444872e-06, "loss": 0.0008, "step": 190800 }, { "epoch": 19.2, "learning_rate": 2.0120044386159587e-06, "loss": 0.0009, "step": 190850 }, { "epoch": 19.21, "learning_rate": 1.999394734187431e-06, "loss": 0.0024, "step": 190900 }, { "epoch": 19.21, "learning_rate": 1.9867850297589025e-06, "loss": 0.0012, "step": 190950 }, { "epoch": 19.22, "learning_rate": 1.9741753253303744e-06, "loss": 0.0011, "step": 191000 }, { "epoch": 19.22, "eval_cer": 0.05737582425793436, "eval_loss": 0.5624056458473206, "eval_runtime": 1308.7488, "eval_samples_per_second": 6.231, "eval_steps_per_second": 0.779, "step": 191000 }, { "epoch": 19.22, "learning_rate": 1.9615656209018464e-06, "loss": 0.0012, "step": 191050 }, { "epoch": 19.23, "learning_rate": 1.948955916473318e-06, "loss": 0.0008, "step": 191100 }, { "epoch": 19.23, "learning_rate": 1.9363462120447897e-06, "loss": 0.0007, "step": 191150 }, { "epoch": 19.24, "learning_rate": 1.9237365076162617e-06, "loss": 0.001, "step": 191200 }, { "epoch": 19.24, "learning_rate": 1.911126803187733e-06, "loss": 0.0008, "step": 191250 }, { "epoch": 19.25, "learning_rate": 1.8985170987592053e-06, "loss": 0.0007, "step": 191300 }, { "epoch": 19.25, "learning_rate": 1.8859073943306772e-06, "loss": 0.001, "step": 191350 }, { "epoch": 19.26, "learning_rate": 1.8732976899021489e-06, "loss": 0.0013, "step": 191400 }, { "epoch": 19.26, "learning_rate": 1.8606879854736206e-06, "loss": 0.0008, "step": 191450 }, { "epoch": 19.27, "learning_rate": 1.8480782810450925e-06, "loss": 0.0012, "step": 191500 }, { "epoch": 19.27, "eval_cer": 0.0572294970703429, "eval_loss": 0.562415599822998, "eval_runtime": 1314.3621, "eval_samples_per_second": 6.205, "eval_steps_per_second": 0.776, "step": 191500 }, { "epoch": 19.27, "learning_rate": 1.8354685766165642e-06, "loss": 0.0008, "step": 191550 }, { "epoch": 19.28, "learning_rate": 1.8228588721880359e-06, "loss": 0.0011, "step": 191600 }, { "epoch": 19.28, "learning_rate": 1.8102491677595075e-06, "loss": 0.0007, "step": 191650 }, { "epoch": 19.29, "learning_rate": 1.7976394633309797e-06, "loss": 0.0008, "step": 191700 }, { "epoch": 19.29, "learning_rate": 1.7850297589024516e-06, "loss": 0.0009, "step": 191750 }, { "epoch": 19.3, "learning_rate": 1.7724200544739233e-06, "loss": 0.0008, "step": 191800 }, { "epoch": 19.3, "learning_rate": 1.759810350045395e-06, "loss": 0.0011, "step": 191850 }, { "epoch": 19.31, "learning_rate": 1.7472006456168669e-06, "loss": 0.001, "step": 191900 }, { "epoch": 19.31, "learning_rate": 1.7345909411883386e-06, "loss": 0.0009, "step": 191950 }, { "epoch": 19.32, "learning_rate": 1.7219812367598103e-06, "loss": 0.0008, "step": 192000 }, { "epoch": 19.32, "eval_cer": 0.05722327038150923, "eval_loss": 0.5627961158752441, "eval_runtime": 1657.2319, "eval_samples_per_second": 4.921, "eval_steps_per_second": 0.615, "step": 192000 }, { "epoch": 19.32, "learning_rate": 1.7093715323312822e-06, "loss": 0.0007, "step": 192050 }, { "epoch": 19.33, "learning_rate": 1.696761827902754e-06, "loss": 0.0009, "step": 192100 }, { "epoch": 19.33, "learning_rate": 1.684152123474226e-06, "loss": 0.0008, "step": 192150 }, { "epoch": 19.34, "learning_rate": 1.6715424190456977e-06, "loss": 0.0011, "step": 192200 }, { "epoch": 19.34, "learning_rate": 1.6589327146171694e-06, "loss": 0.0012, "step": 192250 }, { "epoch": 19.35, "learning_rate": 1.6463230101886413e-06, "loss": 0.0006, "step": 192300 }, { "epoch": 19.35, "learning_rate": 1.633713305760113e-06, "loss": 0.0007, "step": 192350 }, { "epoch": 19.36, "learning_rate": 1.6211036013315847e-06, "loss": 0.0016, "step": 192400 }, { "epoch": 19.37, "learning_rate": 1.6084938969030568e-06, "loss": 0.0009, "step": 192450 }, { "epoch": 19.37, "learning_rate": 1.5958841924745287e-06, "loss": 0.001, "step": 192500 }, { "epoch": 19.37, "eval_cer": 0.05713609673783772, "eval_loss": 0.562631368637085, "eval_runtime": 1655.5987, "eval_samples_per_second": 4.926, "eval_steps_per_second": 0.616, "step": 192500 }, { "epoch": 19.38, "learning_rate": 1.5832744880460004e-06, "loss": 0.0009, "step": 192550 }, { "epoch": 19.38, "learning_rate": 1.570664783617472e-06, "loss": 0.0007, "step": 192600 }, { "epoch": 19.39, "learning_rate": 1.5580550791889438e-06, "loss": 0.0012, "step": 192650 }, { "epoch": 19.39, "learning_rate": 1.5454453747604157e-06, "loss": 0.001, "step": 192700 }, { "epoch": 19.4, "learning_rate": 1.5328356703318876e-06, "loss": 0.0063, "step": 192750 }, { "epoch": 19.4, "learning_rate": 1.5202259659033593e-06, "loss": 0.0008, "step": 192800 }, { "epoch": 19.41, "learning_rate": 1.507616261474831e-06, "loss": 0.0009, "step": 192850 }, { "epoch": 19.41, "learning_rate": 1.4950065570463029e-06, "loss": 0.0012, "step": 192900 }, { "epoch": 19.42, "learning_rate": 1.4823968526177748e-06, "loss": 0.0017, "step": 192950 }, { "epoch": 19.42, "learning_rate": 1.4697871481892465e-06, "loss": 0.0009, "step": 193000 }, { "epoch": 19.42, "eval_cer": 0.05709562326041881, "eval_loss": 0.5627738833427429, "eval_runtime": 1470.9851, "eval_samples_per_second": 5.544, "eval_steps_per_second": 0.693, "step": 193000 }, { "epoch": 19.43, "learning_rate": 1.4571774437607184e-06, "loss": 0.0012, "step": 193050 }, { "epoch": 19.43, "learning_rate": 1.44456773933219e-06, "loss": 0.001, "step": 193100 }, { "epoch": 19.44, "learning_rate": 1.431958034903662e-06, "loss": 0.0008, "step": 193150 }, { "epoch": 19.44, "learning_rate": 1.4193483304751337e-06, "loss": 0.0012, "step": 193200 }, { "epoch": 19.45, "learning_rate": 1.4067386260466056e-06, "loss": 0.001, "step": 193250 }, { "epoch": 19.45, "learning_rate": 1.394381115706648e-06, "loss": 0.001, "step": 193300 }, { "epoch": 19.46, "learning_rate": 1.3817714112781197e-06, "loss": 0.0007, "step": 193350 }, { "epoch": 19.46, "learning_rate": 1.3691617068495916e-06, "loss": 0.0007, "step": 193400 }, { "epoch": 19.47, "learning_rate": 1.3565520024210633e-06, "loss": 0.0007, "step": 193450 }, { "epoch": 19.47, "learning_rate": 1.3439422979925352e-06, "loss": 0.0011, "step": 193500 }, { "epoch": 19.47, "eval_cer": 0.05713609673783772, "eval_loss": 0.5628530383110046, "eval_runtime": 1669.7003, "eval_samples_per_second": 4.884, "eval_steps_per_second": 0.611, "step": 193500 }, { "epoch": 19.48, "learning_rate": 1.3313325935640069e-06, "loss": 0.0008, "step": 193550 }, { "epoch": 19.48, "learning_rate": 1.3187228891354788e-06, "loss": 0.0009, "step": 193600 }, { "epoch": 19.49, "learning_rate": 1.3061131847069505e-06, "loss": 0.0013, "step": 193650 }, { "epoch": 19.49, "learning_rate": 1.2935034802784224e-06, "loss": 0.0017, "step": 193700 }, { "epoch": 19.5, "learning_rate": 1.280893775849894e-06, "loss": 0.0009, "step": 193750 }, { "epoch": 19.5, "learning_rate": 1.268284071421366e-06, "loss": 0.0012, "step": 193800 }, { "epoch": 19.51, "learning_rate": 1.2556743669928377e-06, "loss": 0.0011, "step": 193850 }, { "epoch": 19.51, "learning_rate": 1.2430646625643096e-06, "loss": 0.0013, "step": 193900 }, { "epoch": 19.52, "learning_rate": 1.2304549581357813e-06, "loss": 0.001, "step": 193950 }, { "epoch": 19.52, "learning_rate": 1.2178452537072532e-06, "loss": 0.0008, "step": 194000 }, { "epoch": 19.52, "eval_cer": 0.05709562326041881, "eval_loss": 0.5622166395187378, "eval_runtime": 1646.182, "eval_samples_per_second": 4.954, "eval_steps_per_second": 0.62, "step": 194000 }, { "epoch": 19.53, "learning_rate": 1.2052355492787249e-06, "loss": 0.0009, "step": 194050 }, { "epoch": 19.53, "learning_rate": 1.1926258448501968e-06, "loss": 0.0007, "step": 194100 }, { "epoch": 19.54, "learning_rate": 1.1800161404216685e-06, "loss": 0.0011, "step": 194150 }, { "epoch": 19.54, "learning_rate": 1.1674064359931404e-06, "loss": 0.0011, "step": 194200 }, { "epoch": 19.55, "learning_rate": 1.154796731564612e-06, "loss": 0.0009, "step": 194250 }, { "epoch": 19.55, "learning_rate": 1.142187027136084e-06, "loss": 0.001, "step": 194300 }, { "epoch": 19.56, "learning_rate": 1.1295773227075557e-06, "loss": 0.0012, "step": 194350 }, { "epoch": 19.56, "learning_rate": 1.1169676182790276e-06, "loss": 0.001, "step": 194400 }, { "epoch": 19.57, "learning_rate": 1.1043579138504995e-06, "loss": 0.0012, "step": 194450 }, { "epoch": 19.57, "learning_rate": 1.0917482094219712e-06, "loss": 0.0008, "step": 194500 }, { "epoch": 19.57, "eval_cer": 0.05697420282816207, "eval_loss": 0.5625803470611572, "eval_runtime": 1590.7956, "eval_samples_per_second": 5.126, "eval_steps_per_second": 0.641, "step": 194500 }, { "epoch": 19.58, "learning_rate": 1.079138504993443e-06, "loss": 0.001, "step": 194550 }, { "epoch": 19.58, "learning_rate": 1.0665288005649148e-06, "loss": 0.001, "step": 194600 }, { "epoch": 19.59, "learning_rate": 1.0539190961363867e-06, "loss": 0.0009, "step": 194650 }, { "epoch": 19.59, "learning_rate": 1.0413093917078584e-06, "loss": 0.0012, "step": 194700 }, { "epoch": 19.6, "learning_rate": 1.0286996872793303e-06, "loss": 0.0006, "step": 194750 }, { "epoch": 19.6, "learning_rate": 1.016089982850802e-06, "loss": 0.0008, "step": 194800 }, { "epoch": 19.61, "learning_rate": 1.003480278422274e-06, "loss": 0.001, "step": 194850 }, { "epoch": 19.61, "learning_rate": 9.908705739937456e-07, "loss": 0.0008, "step": 194900 }, { "epoch": 19.62, "learning_rate": 9.782608695652175e-07, "loss": 0.0008, "step": 194950 }, { "epoch": 19.62, "learning_rate": 9.656511651366892e-07, "loss": 0.0009, "step": 195000 }, { "epoch": 19.62, "eval_cer": 0.05706448981625041, "eval_loss": 0.5624039769172668, "eval_runtime": 1354.1662, "eval_samples_per_second": 6.022, "eval_steps_per_second": 0.753, "step": 195000 }, { "epoch": 19.63, "learning_rate": 9.53041460708161e-07, "loss": 0.0009, "step": 195050 }, { "epoch": 19.63, "learning_rate": 9.404317562796328e-07, "loss": 0.001, "step": 195100 }, { "epoch": 19.64, "learning_rate": 9.278220518511047e-07, "loss": 0.0011, "step": 195150 }, { "epoch": 19.64, "learning_rate": 9.152123474225765e-07, "loss": 0.001, "step": 195200 }, { "epoch": 19.65, "learning_rate": 9.026026429940482e-07, "loss": 0.0009, "step": 195250 }, { "epoch": 19.65, "learning_rate": 8.8999293856552e-07, "loss": 0.0007, "step": 195300 }, { "epoch": 19.66, "learning_rate": 8.773832341369919e-07, "loss": 0.0009, "step": 195350 }, { "epoch": 19.66, "learning_rate": 8.647735297084637e-07, "loss": 0.0009, "step": 195400 }, { "epoch": 19.67, "learning_rate": 8.521638252799354e-07, "loss": 0.001, "step": 195450 }, { "epoch": 19.67, "learning_rate": 8.395541208514072e-07, "loss": 0.001, "step": 195500 }, { "epoch": 19.67, "eval_cer": 0.056939956039576835, "eval_loss": 0.5625380873680115, "eval_runtime": 1311.1915, "eval_samples_per_second": 6.22, "eval_steps_per_second": 0.778, "step": 195500 }, { "epoch": 19.68, "learning_rate": 8.269444164228791e-07, "loss": 0.0011, "step": 195550 }, { "epoch": 19.68, "learning_rate": 8.143347119943509e-07, "loss": 0.0016, "step": 195600 }, { "epoch": 19.69, "learning_rate": 8.017250075658226e-07, "loss": 0.001, "step": 195650 }, { "epoch": 19.69, "learning_rate": 7.891153031372944e-07, "loss": 0.0011, "step": 195700 }, { "epoch": 19.7, "learning_rate": 7.765055987087662e-07, "loss": 0.0008, "step": 195750 }, { "epoch": 19.7, "learning_rate": 7.638958942802381e-07, "loss": 0.0007, "step": 195800 }, { "epoch": 19.71, "learning_rate": 7.512861898517098e-07, "loss": 0.001, "step": 195850 }, { "epoch": 19.71, "learning_rate": 7.386764854231817e-07, "loss": 0.0012, "step": 195900 }, { "epoch": 19.72, "learning_rate": 7.260667809946534e-07, "loss": 0.0008, "step": 195950 }, { "epoch": 19.72, "learning_rate": 7.134570765661253e-07, "loss": 0.0011, "step": 196000 }, { "epoch": 19.72, "eval_cer": 0.05679674219640222, "eval_loss": 0.5630409121513367, "eval_runtime": 1314.3616, "eval_samples_per_second": 6.205, "eval_steps_per_second": 0.776, "step": 196000 }, { "epoch": 19.73, "learning_rate": 7.00847372137597e-07, "loss": 0.0011, "step": 196050 }, { "epoch": 19.73, "learning_rate": 6.882376677090689e-07, "loss": 0.0007, "step": 196100 }, { "epoch": 19.74, "learning_rate": 6.756279632805407e-07, "loss": 0.001, "step": 196150 }, { "epoch": 19.74, "learning_rate": 6.630182588520125e-07, "loss": 0.0013, "step": 196200 }, { "epoch": 19.75, "learning_rate": 6.504085544234843e-07, "loss": 0.0008, "step": 196250 }, { "epoch": 19.75, "learning_rate": 6.377988499949561e-07, "loss": 0.0007, "step": 196300 }, { "epoch": 19.76, "learning_rate": 6.25189145566428e-07, "loss": 0.0008, "step": 196350 }, { "epoch": 19.76, "learning_rate": 6.125794411378998e-07, "loss": 0.001, "step": 196400 }, { "epoch": 19.77, "learning_rate": 5.999697367093716e-07, "loss": 0.0008, "step": 196450 }, { "epoch": 19.77, "learning_rate": 5.876122263694139e-07, "loss": 0.0012, "step": 196500 }, { "epoch": 19.77, "eval_cer": 0.05681230891848642, "eval_loss": 0.5628609657287598, "eval_runtime": 1334.2683, "eval_samples_per_second": 6.112, "eval_steps_per_second": 0.764, "step": 196500 }, { "epoch": 19.78, "learning_rate": 5.750025219408857e-07, "loss": 0.0008, "step": 196550 }, { "epoch": 19.78, "learning_rate": 5.623928175123575e-07, "loss": 0.0007, "step": 196600 }, { "epoch": 19.79, "learning_rate": 5.497831130838293e-07, "loss": 0.0011, "step": 196650 }, { "epoch": 19.79, "learning_rate": 5.371734086553011e-07, "loss": 0.0011, "step": 196700 }, { "epoch": 19.8, "learning_rate": 5.245637042267729e-07, "loss": 0.0011, "step": 196750 }, { "epoch": 19.8, "learning_rate": 5.119539997982447e-07, "loss": 0.001, "step": 196800 }, { "epoch": 19.81, "learning_rate": 4.993442953697165e-07, "loss": 0.0009, "step": 196850 }, { "epoch": 19.81, "learning_rate": 4.867345909411883e-07, "loss": 0.0008, "step": 196900 }, { "epoch": 19.82, "learning_rate": 4.741248865126602e-07, "loss": 0.0008, "step": 196950 }, { "epoch": 19.82, "learning_rate": 4.6151518208413204e-07, "loss": 0.0019, "step": 197000 }, { "epoch": 19.82, "eval_cer": 0.05684344236265481, "eval_loss": 0.5629487037658691, "eval_runtime": 1345.3429, "eval_samples_per_second": 6.062, "eval_steps_per_second": 0.758, "step": 197000 }, { "epoch": 19.83, "learning_rate": 4.489054776556038e-07, "loss": 0.001, "step": 197050 }, { "epoch": 19.83, "learning_rate": 4.3629577322707564e-07, "loss": 0.0008, "step": 197100 }, { "epoch": 19.84, "learning_rate": 4.236860687985474e-07, "loss": 0.0008, "step": 197150 }, { "epoch": 19.84, "learning_rate": 4.1107636437001924e-07, "loss": 0.0009, "step": 197200 }, { "epoch": 19.85, "learning_rate": 3.98466659941491e-07, "loss": 0.0007, "step": 197250 }, { "epoch": 19.85, "learning_rate": 3.858569555129628e-07, "loss": 0.001, "step": 197300 }, { "epoch": 19.86, "learning_rate": 3.732472510844346e-07, "loss": 0.0008, "step": 197350 }, { "epoch": 19.86, "learning_rate": 3.606375466559064e-07, "loss": 0.0011, "step": 197400 }, { "epoch": 19.87, "learning_rate": 3.480278422273782e-07, "loss": 0.0009, "step": 197450 }, { "epoch": 19.87, "learning_rate": 3.3541813779885e-07, "loss": 0.001, "step": 197500 }, { "epoch": 19.87, "eval_cer": 0.056865235773572685, "eval_loss": 0.5629896521568298, "eval_runtime": 1451.2536, "eval_samples_per_second": 5.619, "eval_steps_per_second": 0.703, "step": 197500 }, { "epoch": 19.88, "learning_rate": 3.228084333703218e-07, "loss": 0.001, "step": 197550 }, { "epoch": 19.88, "learning_rate": 3.101987289417936e-07, "loss": 0.0018, "step": 197600 }, { "epoch": 19.89, "learning_rate": 2.975890245132654e-07, "loss": 0.0008, "step": 197650 }, { "epoch": 19.89, "learning_rate": 2.849793200847372e-07, "loss": 0.0011, "step": 197700 }, { "epoch": 19.9, "learning_rate": 2.7236961565620905e-07, "loss": 0.0009, "step": 197750 }, { "epoch": 19.9, "learning_rate": 2.5975991122768085e-07, "loss": 0.0013, "step": 197800 }, { "epoch": 19.91, "learning_rate": 2.4715020679915266e-07, "loss": 0.0008, "step": 197850 }, { "epoch": 19.91, "learning_rate": 2.3454050237062446e-07, "loss": 0.0007, "step": 197900 }, { "epoch": 19.92, "learning_rate": 2.2193079794209626e-07, "loss": 0.0008, "step": 197950 }, { "epoch": 19.92, "learning_rate": 2.0932109351356806e-07, "loss": 0.0013, "step": 198000 }, { "epoch": 19.92, "eval_cer": 0.05679362885198538, "eval_loss": 0.5629528164863586, "eval_runtime": 1428.7224, "eval_samples_per_second": 5.708, "eval_steps_per_second": 0.714, "step": 198000 }, { "epoch": 19.93, "learning_rate": 1.9671138908503986e-07, "loss": 0.0007, "step": 198050 }, { "epoch": 19.93, "learning_rate": 1.8410168465651166e-07, "loss": 0.0008, "step": 198100 }, { "epoch": 19.94, "learning_rate": 1.7149198022798346e-07, "loss": 0.0008, "step": 198150 }, { "epoch": 19.94, "learning_rate": 1.588822757994553e-07, "loss": 0.0008, "step": 198200 }, { "epoch": 19.95, "learning_rate": 1.462725713709271e-07, "loss": 0.0011, "step": 198250 }, { "epoch": 19.95, "learning_rate": 1.336628669423989e-07, "loss": 0.0007, "step": 198300 }, { "epoch": 19.96, "learning_rate": 1.210531625138707e-07, "loss": 0.001, "step": 198350 }, { "epoch": 19.96, "learning_rate": 1.0869565217391305e-07, "loss": 0.0008, "step": 198400 }, { "epoch": 19.97, "learning_rate": 9.608594774538485e-08, "loss": 0.0018, "step": 198450 }, { "epoch": 19.97, "learning_rate": 8.347624331685665e-08, "loss": 0.0008, "step": 198500 }, { "epoch": 19.97, "eval_cer": 0.05673447530806543, "eval_loss": 0.5630419850349426, "eval_runtime": 1358.5761, "eval_samples_per_second": 6.003, "eval_steps_per_second": 0.751, "step": 198500 } ], "max_steps": 198760, "num_train_epochs": 20, "total_flos": 4.753306577067194e+21, "trial_name": null, "trial_params": null }