| { | |
| "best_metric": 0.4354272186756134, | |
| "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-bemgen-100f50m-model/checkpoint-1400", | |
| "epoch": 2.6391548365797295, | |
| "eval_steps": 200, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03301419610432486, | |
| "grad_norm": 117.47059631347656, | |
| "learning_rate": 3.8e-07, | |
| "loss": 10.9949, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.06602839220864971, | |
| "grad_norm": 102.7651138305664, | |
| "learning_rate": 8.8e-07, | |
| "loss": 9.4482, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.09904258831297458, | |
| "grad_norm": 64.81011199951172, | |
| "learning_rate": 1.3800000000000001e-06, | |
| "loss": 7.2641, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.13205678441729943, | |
| "grad_norm": 81.74808502197266, | |
| "learning_rate": 1.8800000000000002e-06, | |
| "loss": 5.5236, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1650709805216243, | |
| "grad_norm": 93.43115234375, | |
| "learning_rate": 2.38e-06, | |
| "loss": 4.7369, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.19808517662594916, | |
| "grad_norm": 89.20719909667969, | |
| "learning_rate": 2.88e-06, | |
| "loss": 3.8232, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.231099372730274, | |
| "grad_norm": 64.15678405761719, | |
| "learning_rate": 3.3800000000000007e-06, | |
| "loss": 3.806, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 0.26411356883459886, | |
| "grad_norm": 64.91326904296875, | |
| "learning_rate": 3.88e-06, | |
| "loss": 3.1512, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.26411356883459886, | |
| "eval_loss": 0.8282185792922974, | |
| "eval_runtime": 423.1335, | |
| "eval_samples_per_second": 1.791, | |
| "eval_steps_per_second": 0.896, | |
| "eval_wer": 0.6533637400228051, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.29712776493892373, | |
| "grad_norm": 75.13513946533203, | |
| "learning_rate": 4.38e-06, | |
| "loss": 3.3773, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 0.3301419610432486, | |
| "grad_norm": 62.38654708862305, | |
| "learning_rate": 4.880000000000001e-06, | |
| "loss": 3.3841, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.36315615714757343, | |
| "grad_norm": 49.35546875, | |
| "learning_rate": 5.380000000000001e-06, | |
| "loss": 2.8367, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 0.3961703532518983, | |
| "grad_norm": 77.38804626464844, | |
| "learning_rate": 5.8800000000000005e-06, | |
| "loss": 2.9174, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.4291845493562232, | |
| "grad_norm": 78.50695037841797, | |
| "learning_rate": 6.380000000000001e-06, | |
| "loss": 2.6359, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 0.462198745460548, | |
| "grad_norm": 55.40019607543945, | |
| "learning_rate": 6.88e-06, | |
| "loss": 2.4471, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4952129415648729, | |
| "grad_norm": 55.860809326171875, | |
| "learning_rate": 7.3800000000000005e-06, | |
| "loss": 2.6426, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 0.5282271376691977, | |
| "grad_norm": 51.24520492553711, | |
| "learning_rate": 7.88e-06, | |
| "loss": 2.4008, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5282271376691977, | |
| "eval_loss": 0.6202582716941833, | |
| "eval_runtime": 422.1868, | |
| "eval_samples_per_second": 1.795, | |
| "eval_steps_per_second": 0.898, | |
| "eval_wer": 0.5096921322690992, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5612413337735226, | |
| "grad_norm": 70.62708282470703, | |
| "learning_rate": 8.380000000000001e-06, | |
| "loss": 2.4166, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 0.5942555298778475, | |
| "grad_norm": 54.51411437988281, | |
| "learning_rate": 8.880000000000001e-06, | |
| "loss": 2.7424, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.6272697259821723, | |
| "grad_norm": 51.03986740112305, | |
| "learning_rate": 9.38e-06, | |
| "loss": 2.2441, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 0.6602839220864972, | |
| "grad_norm": 64.73434448242188, | |
| "learning_rate": 9.88e-06, | |
| "loss": 2.1702, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6932981181908221, | |
| "grad_norm": 58.62995147705078, | |
| "learning_rate": 9.957777777777779e-06, | |
| "loss": 2.3904, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 0.7263123142951469, | |
| "grad_norm": 61.52521514892578, | |
| "learning_rate": 9.902222222222223e-06, | |
| "loss": 2.2456, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.7593265103994717, | |
| "grad_norm": 45.40699768066406, | |
| "learning_rate": 9.846666666666668e-06, | |
| "loss": 2.3398, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 0.7923407065037966, | |
| "grad_norm": 52.031795501708984, | |
| "learning_rate": 9.791111111111112e-06, | |
| "loss": 2.2459, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.7923407065037966, | |
| "eval_loss": 0.5506373643875122, | |
| "eval_runtime": 424.3675, | |
| "eval_samples_per_second": 1.786, | |
| "eval_steps_per_second": 0.893, | |
| "eval_wer": 0.4643671607753706, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.8253549026081215, | |
| "grad_norm": 51.51828384399414, | |
| "learning_rate": 9.735555555555556e-06, | |
| "loss": 2.1322, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 0.8583690987124464, | |
| "grad_norm": 47.97977066040039, | |
| "learning_rate": 9.68e-06, | |
| "loss": 2.3046, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.8913832948167713, | |
| "grad_norm": 47.63246154785156, | |
| "learning_rate": 9.624444444444445e-06, | |
| "loss": 2.0285, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 0.924397490921096, | |
| "grad_norm": 36.78068161010742, | |
| "learning_rate": 9.56888888888889e-06, | |
| "loss": 1.9158, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.9574116870254209, | |
| "grad_norm": 57.10271453857422, | |
| "learning_rate": 9.513333333333334e-06, | |
| "loss": 2.034, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 0.9904258831297458, | |
| "grad_norm": 35.37094497680664, | |
| "learning_rate": 9.457777777777778e-06, | |
| "loss": 1.92, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.0224496533509408, | |
| "grad_norm": 59.17182159423828, | |
| "learning_rate": 9.402222222222222e-06, | |
| "loss": 1.5202, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 1.0554638494552657, | |
| "grad_norm": 48.809059143066406, | |
| "learning_rate": 9.346666666666666e-06, | |
| "loss": 1.3319, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.0554638494552657, | |
| "eval_loss": 0.5028851628303528, | |
| "eval_runtime": 423.0433, | |
| "eval_samples_per_second": 1.792, | |
| "eval_steps_per_second": 0.896, | |
| "eval_wer": 0.4016533637400228, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.0884780455595906, | |
| "grad_norm": 59.115562438964844, | |
| "learning_rate": 9.291111111111112e-06, | |
| "loss": 1.2718, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 1.1214922416639155, | |
| "grad_norm": 22.62116050720215, | |
| "learning_rate": 9.235555555555556e-06, | |
| "loss": 1.4598, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.1545064377682404, | |
| "grad_norm": 37.58869934082031, | |
| "learning_rate": 9.180000000000002e-06, | |
| "loss": 1.33, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 1.1875206338725652, | |
| "grad_norm": 41.18476867675781, | |
| "learning_rate": 9.124444444444444e-06, | |
| "loss": 1.4802, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.2205348299768901, | |
| "grad_norm": 34.091915130615234, | |
| "learning_rate": 9.06888888888889e-06, | |
| "loss": 1.325, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 1.253549026081215, | |
| "grad_norm": 43.94548416137695, | |
| "learning_rate": 9.013333333333334e-06, | |
| "loss": 1.3758, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.2865632221855399, | |
| "grad_norm": 50.9454231262207, | |
| "learning_rate": 8.957777777777778e-06, | |
| "loss": 1.4197, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 1.3195774182898647, | |
| "grad_norm": 39.400634765625, | |
| "learning_rate": 8.902222222222224e-06, | |
| "loss": 1.5588, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.3195774182898647, | |
| "eval_loss": 0.4675042927265167, | |
| "eval_runtime": 426.8283, | |
| "eval_samples_per_second": 1.776, | |
| "eval_steps_per_second": 0.888, | |
| "eval_wer": 0.3896807297605473, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.3525916143941896, | |
| "grad_norm": 64.05392456054688, | |
| "learning_rate": 8.846666666666668e-06, | |
| "loss": 1.4914, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 1.3856058104985143, | |
| "grad_norm": 38.36800765991211, | |
| "learning_rate": 8.791111111111112e-06, | |
| "loss": 1.3193, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.4186200066028392, | |
| "grad_norm": 56.90019226074219, | |
| "learning_rate": 8.735555555555556e-06, | |
| "loss": 1.3192, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 1.451634202707164, | |
| "grad_norm": 41.730682373046875, | |
| "learning_rate": 8.68e-06, | |
| "loss": 1.4852, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.484648398811489, | |
| "grad_norm": 37.495849609375, | |
| "learning_rate": 8.624444444444446e-06, | |
| "loss": 1.244, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 1.5176625949158138, | |
| "grad_norm": 34.61660385131836, | |
| "learning_rate": 8.56888888888889e-06, | |
| "loss": 1.3055, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.5506767910201387, | |
| "grad_norm": 32.87508773803711, | |
| "learning_rate": 8.513333333333335e-06, | |
| "loss": 1.4434, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 1.5836909871244635, | |
| "grad_norm": 43.57270812988281, | |
| "learning_rate": 8.457777777777778e-06, | |
| "loss": 1.2908, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.5836909871244635, | |
| "eval_loss": 0.45341047644615173, | |
| "eval_runtime": 422.1999, | |
| "eval_samples_per_second": 1.795, | |
| "eval_steps_per_second": 0.898, | |
| "eval_wer": 0.3727194982896237, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.6167051832287884, | |
| "grad_norm": 46.83863830566406, | |
| "learning_rate": 8.402222222222223e-06, | |
| "loss": 1.3848, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 1.649719379333113, | |
| "grad_norm": 46.93707275390625, | |
| "learning_rate": 8.346666666666668e-06, | |
| "loss": 1.2536, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.682733575437438, | |
| "grad_norm": 44.47553253173828, | |
| "learning_rate": 8.291111111111112e-06, | |
| "loss": 1.3358, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 1.7157477715417628, | |
| "grad_norm": 35.15113830566406, | |
| "learning_rate": 8.235555555555557e-06, | |
| "loss": 1.1214, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.7487619676460877, | |
| "grad_norm": 27.538331985473633, | |
| "learning_rate": 8.18e-06, | |
| "loss": 1.097, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 1.7817761637504126, | |
| "grad_norm": 35.05278015136719, | |
| "learning_rate": 8.124444444444445e-06, | |
| "loss": 1.1746, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.8147903598547375, | |
| "grad_norm": 50.673095703125, | |
| "learning_rate": 8.06888888888889e-06, | |
| "loss": 1.397, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 1.8478045559590623, | |
| "grad_norm": 38.40850067138672, | |
| "learning_rate": 8.013333333333333e-06, | |
| "loss": 1.4258, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.8478045559590623, | |
| "eval_loss": 0.4354272186756134, | |
| "eval_runtime": 431.6561, | |
| "eval_samples_per_second": 1.756, | |
| "eval_steps_per_second": 0.878, | |
| "eval_wer": 0.38982326111744586, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.8808187520633872, | |
| "grad_norm": 27.05866050720215, | |
| "learning_rate": 7.957777777777779e-06, | |
| "loss": 1.4042, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 1.913832948167712, | |
| "grad_norm": 36.17206573486328, | |
| "learning_rate": 7.902222222222223e-06, | |
| "loss": 1.0702, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.946847144272037, | |
| "grad_norm": 40.80569076538086, | |
| "learning_rate": 7.846666666666667e-06, | |
| "loss": 1.2552, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 1.9798613403763619, | |
| "grad_norm": 34.74424362182617, | |
| "learning_rate": 7.791111111111111e-06, | |
| "loss": 1.349, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.011885110597557, | |
| "grad_norm": 18.330013275146484, | |
| "learning_rate": 7.735555555555557e-06, | |
| "loss": 1.0107, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 2.0448993067018817, | |
| "grad_norm": 20.316329956054688, | |
| "learning_rate": 7.680000000000001e-06, | |
| "loss": 0.6744, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 2.0779135028062066, | |
| "grad_norm": 56.3316764831543, | |
| "learning_rate": 7.624444444444445e-06, | |
| "loss": 0.6948, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 2.1109276989105314, | |
| "grad_norm": 28.286052703857422, | |
| "learning_rate": 7.56888888888889e-06, | |
| "loss": 0.6383, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.1109276989105314, | |
| "eval_loss": 0.44802892208099365, | |
| "eval_runtime": 429.0328, | |
| "eval_samples_per_second": 1.767, | |
| "eval_steps_per_second": 0.883, | |
| "eval_wer": 0.3600342075256556, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.1439418950148563, | |
| "grad_norm": 33.71493148803711, | |
| "learning_rate": 7.513333333333334e-06, | |
| "loss": 0.5558, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 2.176956091119181, | |
| "grad_norm": 29.385910034179688, | |
| "learning_rate": 7.457777777777778e-06, | |
| "loss": 0.6777, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 2.209970287223506, | |
| "grad_norm": 20.755077362060547, | |
| "learning_rate": 7.402222222222223e-06, | |
| "loss": 0.6332, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 2.242984483327831, | |
| "grad_norm": 37.17219543457031, | |
| "learning_rate": 7.346666666666668e-06, | |
| "loss": 0.6242, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.275998679432156, | |
| "grad_norm": 37.963409423828125, | |
| "learning_rate": 7.291111111111112e-06, | |
| "loss": 0.7322, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 2.3090128755364807, | |
| "grad_norm": 25.863115310668945, | |
| "learning_rate": 7.235555555555556e-06, | |
| "loss": 0.697, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 2.3420270716408056, | |
| "grad_norm": 25.60062026977539, | |
| "learning_rate": 7.180000000000001e-06, | |
| "loss": 0.6266, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 2.3750412677451305, | |
| "grad_norm": 20.695463180541992, | |
| "learning_rate": 7.124444444444445e-06, | |
| "loss": 0.6079, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.3750412677451305, | |
| "eval_loss": 0.4444292187690735, | |
| "eval_runtime": 421.917, | |
| "eval_samples_per_second": 1.797, | |
| "eval_steps_per_second": 0.898, | |
| "eval_wer": 0.3482041049030787, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.4080554638494553, | |
| "grad_norm": 26.363622665405273, | |
| "learning_rate": 7.06888888888889e-06, | |
| "loss": 0.663, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 2.4410696599537802, | |
| "grad_norm": 31.212791442871094, | |
| "learning_rate": 7.0133333333333345e-06, | |
| "loss": 0.655, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 2.474083856058105, | |
| "grad_norm": 39.38019561767578, | |
| "learning_rate": 6.9577777777777785e-06, | |
| "loss": 0.7364, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 2.50709805216243, | |
| "grad_norm": 26.713680267333984, | |
| "learning_rate": 6.902222222222223e-06, | |
| "loss": 0.647, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.540112248266755, | |
| "grad_norm": 24.708005905151367, | |
| "learning_rate": 6.846666666666667e-06, | |
| "loss": 0.6818, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 2.5731264443710797, | |
| "grad_norm": 22.732967376708984, | |
| "learning_rate": 6.7911111111111115e-06, | |
| "loss": 0.6711, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.606140640475404, | |
| "grad_norm": 23.73015022277832, | |
| "learning_rate": 6.735555555555556e-06, | |
| "loss": 0.6492, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 2.6391548365797295, | |
| "grad_norm": 30.428665161132812, | |
| "learning_rate": 6.680000000000001e-06, | |
| "loss": 0.5709, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.6391548365797295, | |
| "eval_loss": 0.4367344081401825, | |
| "eval_runtime": 423.1943, | |
| "eval_samples_per_second": 1.791, | |
| "eval_steps_per_second": 0.896, | |
| "eval_wer": 0.3405074116305587, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.6391548365797295, | |
| "step": 2000, | |
| "total_flos": 1.631539005161472e+19, | |
| "train_loss": 1.9525959782600404, | |
| "train_runtime": 8300.7195, | |
| "train_samples_per_second": 4.819, | |
| "train_steps_per_second": 0.602 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 7, | |
| "save_steps": 200, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.631539005161472e+19, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |