| { | |
| "best_metric": 1.5300506353378296, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-150", | |
| "epoch": 0.07437709185570844, | |
| "eval_steps": 50, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0004958472790380563, | |
| "grad_norm": 2.3338756561279297, | |
| "learning_rate": 1e-05, | |
| "loss": 2.2832, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0004958472790380563, | |
| "eval_loss": 2.5524938106536865, | |
| "eval_runtime": 100.9356, | |
| "eval_samples_per_second": 33.655, | |
| "eval_steps_per_second": 8.421, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0009916945580761127, | |
| "grad_norm": 2.2431490421295166, | |
| "learning_rate": 2e-05, | |
| "loss": 2.0928, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0014875418371141688, | |
| "grad_norm": 2.2695679664611816, | |
| "learning_rate": 3e-05, | |
| "loss": 2.1083, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0019833891161522253, | |
| "grad_norm": 2.306091547012329, | |
| "learning_rate": 4e-05, | |
| "loss": 2.2178, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0024792363951902812, | |
| "grad_norm": 2.505962610244751, | |
| "learning_rate": 5e-05, | |
| "loss": 2.285, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0029750836742283376, | |
| "grad_norm": 2.5917468070983887, | |
| "learning_rate": 6e-05, | |
| "loss": 2.1684, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.003470930953266394, | |
| "grad_norm": 2.413541078567505, | |
| "learning_rate": 7e-05, | |
| "loss": 1.8237, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.003966778232304451, | |
| "grad_norm": 1.8782087564468384, | |
| "learning_rate": 8e-05, | |
| "loss": 1.7509, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.004462625511342506, | |
| "grad_norm": 2.012852430343628, | |
| "learning_rate": 9e-05, | |
| "loss": 1.8674, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.0049584727903805624, | |
| "grad_norm": 1.869946837425232, | |
| "learning_rate": 0.0001, | |
| "loss": 1.5963, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.005454320069418619, | |
| "grad_norm": 2.0942704677581787, | |
| "learning_rate": 9.999316524962345e-05, | |
| "loss": 1.6315, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.005950167348456675, | |
| "grad_norm": 1.4826569557189941, | |
| "learning_rate": 9.997266286704631e-05, | |
| "loss": 1.4265, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.006446014627494731, | |
| "grad_norm": 1.4447063207626343, | |
| "learning_rate": 9.993849845741524e-05, | |
| "loss": 1.5016, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.006941861906532788, | |
| "grad_norm": 1.4674795866012573, | |
| "learning_rate": 9.989068136093873e-05, | |
| "loss": 1.3457, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.007437709185570844, | |
| "grad_norm": 1.3208996057510376, | |
| "learning_rate": 9.98292246503335e-05, | |
| "loss": 1.5108, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.007933556464608901, | |
| "grad_norm": 1.334173321723938, | |
| "learning_rate": 9.975414512725057e-05, | |
| "loss": 1.4771, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.008429403743646957, | |
| "grad_norm": 1.3226054906845093, | |
| "learning_rate": 9.966546331768191e-05, | |
| "loss": 1.489, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.008925251022685012, | |
| "grad_norm": 1.2961368560791016, | |
| "learning_rate": 9.956320346634876e-05, | |
| "loss": 1.7152, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.00942109830172307, | |
| "grad_norm": 1.3078362941741943, | |
| "learning_rate": 9.944739353007344e-05, | |
| "loss": 1.4121, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.009916945580761125, | |
| "grad_norm": 1.2526671886444092, | |
| "learning_rate": 9.931806517013612e-05, | |
| "loss": 1.4345, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.010412792859799182, | |
| "grad_norm": 1.279798984527588, | |
| "learning_rate": 9.917525374361912e-05, | |
| "loss": 1.5639, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.010908640138837238, | |
| "grad_norm": 1.2249804735183716, | |
| "learning_rate": 9.901899829374047e-05, | |
| "loss": 1.6111, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.011404487417875295, | |
| "grad_norm": 1.1494619846343994, | |
| "learning_rate": 9.884934153917997e-05, | |
| "loss": 1.5668, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.01190033469691335, | |
| "grad_norm": 1.080565333366394, | |
| "learning_rate": 9.86663298624003e-05, | |
| "loss": 1.5539, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.012396181975951407, | |
| "grad_norm": 1.2847322225570679, | |
| "learning_rate": 9.847001329696653e-05, | |
| "loss": 1.6474, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.012892029254989463, | |
| "grad_norm": 1.238361120223999, | |
| "learning_rate": 9.826044551386744e-05, | |
| "loss": 1.6814, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.01338787653402752, | |
| "grad_norm": 1.235642910003662, | |
| "learning_rate": 9.803768380684242e-05, | |
| "loss": 1.5172, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.013883723813065576, | |
| "grad_norm": 1.1233747005462646, | |
| "learning_rate": 9.780178907671789e-05, | |
| "loss": 1.4895, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.014379571092103633, | |
| "grad_norm": 1.1913343667984009, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 1.6806, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.014875418371141688, | |
| "grad_norm": 1.0970852375030518, | |
| "learning_rate": 9.729086208503174e-05, | |
| "loss": 1.5504, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.015371265650179745, | |
| "grad_norm": 1.051615595817566, | |
| "learning_rate": 9.701596950580806e-05, | |
| "loss": 1.512, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.015867112929217803, | |
| "grad_norm": 1.088776707649231, | |
| "learning_rate": 9.672822322997305e-05, | |
| "loss": 1.5421, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.016362960208255856, | |
| "grad_norm": 1.1589713096618652, | |
| "learning_rate": 9.642770192448536e-05, | |
| "loss": 1.5206, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.016858807487293914, | |
| "grad_norm": 1.24736750125885, | |
| "learning_rate": 9.611448774886924e-05, | |
| "loss": 1.65, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.01735465476633197, | |
| "grad_norm": 1.1095153093338013, | |
| "learning_rate": 9.578866633275288e-05, | |
| "loss": 1.4902, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.017850502045370024, | |
| "grad_norm": 1.0437101125717163, | |
| "learning_rate": 9.545032675245813e-05, | |
| "loss": 1.5521, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.01834634932440808, | |
| "grad_norm": 1.1120080947875977, | |
| "learning_rate": 9.509956150664796e-05, | |
| "loss": 1.5881, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.01884219660344614, | |
| "grad_norm": 1.1262872219085693, | |
| "learning_rate": 9.473646649103818e-05, | |
| "loss": 1.452, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.019338043882484196, | |
| "grad_norm": 1.0746490955352783, | |
| "learning_rate": 9.43611409721806e-05, | |
| "loss": 1.4759, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.01983389116152225, | |
| "grad_norm": 1.2698118686676025, | |
| "learning_rate": 9.397368756032445e-05, | |
| "loss": 1.6215, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.020329738440560307, | |
| "grad_norm": 1.0959280729293823, | |
| "learning_rate": 9.357421218136386e-05, | |
| "loss": 1.4456, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.020825585719598364, | |
| "grad_norm": 1.1139137744903564, | |
| "learning_rate": 9.316282404787871e-05, | |
| "loss": 1.6712, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.02132143299863642, | |
| "grad_norm": 0.9761843085289001, | |
| "learning_rate": 9.273963562927695e-05, | |
| "loss": 1.4481, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.021817280277674475, | |
| "grad_norm": 1.285643458366394, | |
| "learning_rate": 9.230476262104677e-05, | |
| "loss": 1.7598, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.022313127556712532, | |
| "grad_norm": 1.230698823928833, | |
| "learning_rate": 9.185832391312644e-05, | |
| "loss": 1.7332, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.02280897483575059, | |
| "grad_norm": 1.2132545709609985, | |
| "learning_rate": 9.140044155740101e-05, | |
| "loss": 1.6199, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.023304822114788647, | |
| "grad_norm": 1.3424148559570312, | |
| "learning_rate": 9.093124073433463e-05, | |
| "loss": 1.8396, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.0238006693938267, | |
| "grad_norm": 1.2430864572525024, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 1.6328, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.024296516672864758, | |
| "grad_norm": 1.2584593296051025, | |
| "learning_rate": 8.995939984474624e-05, | |
| "loss": 1.6906, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.024792363951902815, | |
| "grad_norm": 1.4835307598114014, | |
| "learning_rate": 8.945702546981969e-05, | |
| "loss": 1.8958, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.024792363951902815, | |
| "eval_loss": 1.5631545782089233, | |
| "eval_runtime": 101.3122, | |
| "eval_samples_per_second": 33.53, | |
| "eval_steps_per_second": 8.39, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.02528821123094087, | |
| "grad_norm": 1.028480887413025, | |
| "learning_rate": 8.894386393810563e-05, | |
| "loss": 1.4816, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.025784058509978926, | |
| "grad_norm": 1.053328514099121, | |
| "learning_rate": 8.842005554284296e-05, | |
| "loss": 1.3263, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.026279905789016983, | |
| "grad_norm": 1.0090606212615967, | |
| "learning_rate": 8.788574348801675e-05, | |
| "loss": 1.322, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.02677575306805504, | |
| "grad_norm": 1.1087706089019775, | |
| "learning_rate": 8.73410738492077e-05, | |
| "loss": 1.2652, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.027271600347093094, | |
| "grad_norm": 0.9264365434646606, | |
| "learning_rate": 8.678619553365659e-05, | |
| "loss": 1.3602, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.02776744762613115, | |
| "grad_norm": 0.9596973657608032, | |
| "learning_rate": 8.622126023955446e-05, | |
| "loss": 1.4291, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.028263294905169208, | |
| "grad_norm": 1.0175056457519531, | |
| "learning_rate": 8.564642241456986e-05, | |
| "loss": 1.4, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.028759142184207265, | |
| "grad_norm": 1.0313339233398438, | |
| "learning_rate": 8.506183921362443e-05, | |
| "loss": 1.579, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.02925498946324532, | |
| "grad_norm": 1.0789611339569092, | |
| "learning_rate": 8.44676704559283e-05, | |
| "loss": 1.4908, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.029750836742283376, | |
| "grad_norm": 0.9643576741218567, | |
| "learning_rate": 8.386407858128706e-05, | |
| "loss": 1.4567, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.030246684021321434, | |
| "grad_norm": 0.979160487651825, | |
| "learning_rate": 8.32512286056924e-05, | |
| "loss": 1.3803, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.03074253130035949, | |
| "grad_norm": 1.0837715864181519, | |
| "learning_rate": 8.262928807620843e-05, | |
| "loss": 1.5957, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.031238378579397545, | |
| "grad_norm": 1.090160846710205, | |
| "learning_rate": 8.199842702516583e-05, | |
| "loss": 1.4779, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.031734225858435605, | |
| "grad_norm": 1.073777675628662, | |
| "learning_rate": 8.135881792367686e-05, | |
| "loss": 1.6611, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.032230073137473655, | |
| "grad_norm": 1.0346155166625977, | |
| "learning_rate": 8.07106356344834e-05, | |
| "loss": 1.5201, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.03272592041651171, | |
| "grad_norm": 1.0693767070770264, | |
| "learning_rate": 8.005405736415126e-05, | |
| "loss": 1.4886, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.03322176769554977, | |
| "grad_norm": 0.9293779730796814, | |
| "learning_rate": 7.938926261462366e-05, | |
| "loss": 1.3359, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.03371761497458783, | |
| "grad_norm": 0.9526183009147644, | |
| "learning_rate": 7.871643313414718e-05, | |
| "loss": 1.5865, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.034213462253625884, | |
| "grad_norm": 0.9830014705657959, | |
| "learning_rate": 7.803575286758364e-05, | |
| "loss": 1.3337, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.03470930953266394, | |
| "grad_norm": 1.031559705734253, | |
| "learning_rate": 7.734740790612136e-05, | |
| "loss": 1.6285, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.035205156811702, | |
| "grad_norm": 1.0397356748580933, | |
| "learning_rate": 7.66515864363997e-05, | |
| "loss": 1.5587, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.03570100409074005, | |
| "grad_norm": 1.0040336847305298, | |
| "learning_rate": 7.594847868906076e-05, | |
| "loss": 1.4988, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.036196851369778106, | |
| "grad_norm": 1.0021733045578003, | |
| "learning_rate": 7.52382768867422e-05, | |
| "loss": 1.5555, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.03669269864881616, | |
| "grad_norm": 0.9916362762451172, | |
| "learning_rate": 7.452117519152542e-05, | |
| "loss": 1.4792, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.03718854592785422, | |
| "grad_norm": 1.0307399034500122, | |
| "learning_rate": 7.379736965185368e-05, | |
| "loss": 1.5331, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.03768439320689228, | |
| "grad_norm": 0.9940649271011353, | |
| "learning_rate": 7.30670581489344e-05, | |
| "loss": 1.4171, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 0.038180240485930335, | |
| "grad_norm": 0.9914387464523315, | |
| "learning_rate": 7.233044034264034e-05, | |
| "loss": 1.4668, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 0.03867608776496839, | |
| "grad_norm": 1.0079885721206665, | |
| "learning_rate": 7.158771761692464e-05, | |
| "loss": 1.5295, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 0.03917193504400645, | |
| "grad_norm": 0.917401909828186, | |
| "learning_rate": 7.083909302476453e-05, | |
| "loss": 1.3983, | |
| "step": 79 | |
| }, | |
| { | |
| "epoch": 0.0396677823230445, | |
| "grad_norm": 1.0264183282852173, | |
| "learning_rate": 7.008477123264848e-05, | |
| "loss": 1.4092, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.04016362960208256, | |
| "grad_norm": 1.005240797996521, | |
| "learning_rate": 6.932495846462261e-05, | |
| "loss": 1.5231, | |
| "step": 81 | |
| }, | |
| { | |
| "epoch": 0.040659476881120614, | |
| "grad_norm": 1.0037102699279785, | |
| "learning_rate": 6.855986244591104e-05, | |
| "loss": 1.495, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 0.04115532416015867, | |
| "grad_norm": 1.193308711051941, | |
| "learning_rate": 6.778969234612584e-05, | |
| "loss": 1.7662, | |
| "step": 83 | |
| }, | |
| { | |
| "epoch": 0.04165117143919673, | |
| "grad_norm": 1.1065670251846313, | |
| "learning_rate": 6.701465872208216e-05, | |
| "loss": 1.6595, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 0.042147018718234786, | |
| "grad_norm": 1.0933548212051392, | |
| "learning_rate": 6.623497346023418e-05, | |
| "loss": 1.4608, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.04264286599727284, | |
| "grad_norm": 1.045641303062439, | |
| "learning_rate": 6.545084971874738e-05, | |
| "loss": 1.4828, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 0.04313871327631089, | |
| "grad_norm": 1.165783405303955, | |
| "learning_rate": 6.466250186922325e-05, | |
| "loss": 1.7687, | |
| "step": 87 | |
| }, | |
| { | |
| "epoch": 0.04363456055534895, | |
| "grad_norm": 1.4017337560653687, | |
| "learning_rate": 6.387014543809223e-05, | |
| "loss": 1.7701, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 0.04413040783438701, | |
| "grad_norm": 1.0517154932022095, | |
| "learning_rate": 6.307399704769099e-05, | |
| "loss": 1.7043, | |
| "step": 89 | |
| }, | |
| { | |
| "epoch": 0.044626255113425065, | |
| "grad_norm": 1.078675389289856, | |
| "learning_rate": 6.227427435703997e-05, | |
| "loss": 1.5318, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.04512210239246312, | |
| "grad_norm": 1.0338232517242432, | |
| "learning_rate": 6.147119600233758e-05, | |
| "loss": 1.4883, | |
| "step": 91 | |
| }, | |
| { | |
| "epoch": 0.04561794967150118, | |
| "grad_norm": 1.039003610610962, | |
| "learning_rate": 6.066498153718735e-05, | |
| "loss": 1.5021, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 0.046113796950539236, | |
| "grad_norm": 1.154383897781372, | |
| "learning_rate": 5.985585137257401e-05, | |
| "loss": 1.7616, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 0.04660964422957729, | |
| "grad_norm": 1.1290744543075562, | |
| "learning_rate": 5.90440267166055e-05, | |
| "loss": 1.6727, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 0.047105491508615344, | |
| "grad_norm": 1.129178762435913, | |
| "learning_rate": 5.8229729514036705e-05, | |
| "loss": 1.5239, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.0476013387876534, | |
| "grad_norm": 1.0544580221176147, | |
| "learning_rate": 5.74131823855921e-05, | |
| "loss": 1.6728, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 0.04809718606669146, | |
| "grad_norm": 1.2535539865493774, | |
| "learning_rate": 5.6594608567103456e-05, | |
| "loss": 1.7839, | |
| "step": 97 | |
| }, | |
| { | |
| "epoch": 0.048593033345729515, | |
| "grad_norm": 1.105029821395874, | |
| "learning_rate": 5.577423184847932e-05, | |
| "loss": 1.6778, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 0.04908888062476757, | |
| "grad_norm": 1.0066592693328857, | |
| "learning_rate": 5.495227651252315e-05, | |
| "loss": 1.5587, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 0.04958472790380563, | |
| "grad_norm": 1.2827523946762085, | |
| "learning_rate": 5.4128967273616625e-05, | |
| "loss": 1.7548, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04958472790380563, | |
| "eval_loss": 1.5408598184585571, | |
| "eval_runtime": 101.3458, | |
| "eval_samples_per_second": 33.519, | |
| "eval_steps_per_second": 8.387, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.05008057518284369, | |
| "grad_norm": 0.9162317514419556, | |
| "learning_rate": 5.330452921628497e-05, | |
| "loss": 1.4666, | |
| "step": 101 | |
| }, | |
| { | |
| "epoch": 0.05057642246188174, | |
| "grad_norm": 0.9027171730995178, | |
| "learning_rate": 5.247918773366112e-05, | |
| "loss": 1.4267, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 0.051072269740919794, | |
| "grad_norm": 0.8656598329544067, | |
| "learning_rate": 5.165316846586541e-05, | |
| "loss": 1.2987, | |
| "step": 103 | |
| }, | |
| { | |
| "epoch": 0.05156811701995785, | |
| "grad_norm": 0.8837535381317139, | |
| "learning_rate": 5.0826697238317935e-05, | |
| "loss": 1.33, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 0.05206396429899591, | |
| "grad_norm": 1.0209332704544067, | |
| "learning_rate": 5e-05, | |
| "loss": 1.4653, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.052559811578033966, | |
| "grad_norm": 0.883824348449707, | |
| "learning_rate": 4.917330276168208e-05, | |
| "loss": 1.3663, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 0.05305565885707202, | |
| "grad_norm": 1.0715792179107666, | |
| "learning_rate": 4.834683153413459e-05, | |
| "loss": 1.445, | |
| "step": 107 | |
| }, | |
| { | |
| "epoch": 0.05355150613611008, | |
| "grad_norm": 0.926857054233551, | |
| "learning_rate": 4.7520812266338885e-05, | |
| "loss": 1.3781, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 0.05404735341514814, | |
| "grad_norm": 1.0013282299041748, | |
| "learning_rate": 4.669547078371504e-05, | |
| "loss": 1.6475, | |
| "step": 109 | |
| }, | |
| { | |
| "epoch": 0.05454320069418619, | |
| "grad_norm": 0.9990188479423523, | |
| "learning_rate": 4.5871032726383386e-05, | |
| "loss": 1.6305, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.055039047973224245, | |
| "grad_norm": 1.0079963207244873, | |
| "learning_rate": 4.504772348747687e-05, | |
| "loss": 1.3203, | |
| "step": 111 | |
| }, | |
| { | |
| "epoch": 0.0555348952522623, | |
| "grad_norm": 0.9960539937019348, | |
| "learning_rate": 4.4225768151520694e-05, | |
| "loss": 1.4883, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 0.05603074253130036, | |
| "grad_norm": 0.949286699295044, | |
| "learning_rate": 4.3405391432896555e-05, | |
| "loss": 1.3312, | |
| "step": 113 | |
| }, | |
| { | |
| "epoch": 0.056526589810338417, | |
| "grad_norm": 1.0726183652877808, | |
| "learning_rate": 4.2586817614407895e-05, | |
| "loss": 1.6161, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 0.057022437089376474, | |
| "grad_norm": 0.8714136481285095, | |
| "learning_rate": 4.17702704859633e-05, | |
| "loss": 1.3989, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.05751828436841453, | |
| "grad_norm": 0.9169184565544128, | |
| "learning_rate": 4.095597328339452e-05, | |
| "loss": 1.3558, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 0.05801413164745258, | |
| "grad_norm": 0.9536529183387756, | |
| "learning_rate": 4.0144148627425993e-05, | |
| "loss": 1.488, | |
| "step": 117 | |
| }, | |
| { | |
| "epoch": 0.05850997892649064, | |
| "grad_norm": 0.9956008791923523, | |
| "learning_rate": 3.933501846281267e-05, | |
| "loss": 1.4758, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 0.059005826205528696, | |
| "grad_norm": 1.021716594696045, | |
| "learning_rate": 3.852880399766243e-05, | |
| "loss": 1.6887, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 0.05950167348456675, | |
| "grad_norm": 0.9416800141334534, | |
| "learning_rate": 3.772572564296005e-05, | |
| "loss": 1.4239, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.05999752076360481, | |
| "grad_norm": 0.9313503503799438, | |
| "learning_rate": 3.6926002952309016e-05, | |
| "loss": 1.496, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 0.06049336804264287, | |
| "grad_norm": 1.0198074579238892, | |
| "learning_rate": 3.612985456190778e-05, | |
| "loss": 1.6401, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 0.060989215321680924, | |
| "grad_norm": 1.1576324701309204, | |
| "learning_rate": 3.533749813077677e-05, | |
| "loss": 1.597, | |
| "step": 123 | |
| }, | |
| { | |
| "epoch": 0.06148506260071898, | |
| "grad_norm": 1.1293891668319702, | |
| "learning_rate": 3.4549150281252636e-05, | |
| "loss": 1.4957, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 0.06198090987975703, | |
| "grad_norm": 1.1353529691696167, | |
| "learning_rate": 3.3765026539765834e-05, | |
| "loss": 1.7044, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.06247675715879509, | |
| "grad_norm": 0.9573756456375122, | |
| "learning_rate": 3.298534127791785e-05, | |
| "loss": 1.4988, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 0.06297260443783315, | |
| "grad_norm": 0.984106183052063, | |
| "learning_rate": 3.221030765387417e-05, | |
| "loss": 1.4513, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 0.06346845171687121, | |
| "grad_norm": 0.909013032913208, | |
| "learning_rate": 3.144013755408895e-05, | |
| "loss": 1.4702, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 0.06396429899590926, | |
| "grad_norm": 0.8617895841598511, | |
| "learning_rate": 3.0675041535377405e-05, | |
| "loss": 1.2816, | |
| "step": 129 | |
| }, | |
| { | |
| "epoch": 0.06446014627494731, | |
| "grad_norm": 1.0924761295318604, | |
| "learning_rate": 2.991522876735154e-05, | |
| "loss": 1.5841, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.06495599355398537, | |
| "grad_norm": 1.0284861326217651, | |
| "learning_rate": 2.916090697523549e-05, | |
| "loss": 1.5327, | |
| "step": 131 | |
| }, | |
| { | |
| "epoch": 0.06545184083302343, | |
| "grad_norm": 0.9654609560966492, | |
| "learning_rate": 2.8412282383075363e-05, | |
| "loss": 1.3829, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 0.06594768811206149, | |
| "grad_norm": 1.027591347694397, | |
| "learning_rate": 2.766955965735968e-05, | |
| "loss": 1.4823, | |
| "step": 133 | |
| }, | |
| { | |
| "epoch": 0.06644353539109954, | |
| "grad_norm": 0.9814398288726807, | |
| "learning_rate": 2.693294185106562e-05, | |
| "loss": 1.5405, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 0.0669393826701376, | |
| "grad_norm": 1.0590242147445679, | |
| "learning_rate": 2.6202630348146324e-05, | |
| "loss": 1.4063, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 0.06743522994917565, | |
| "grad_norm": 1.008570671081543, | |
| "learning_rate": 2.547882480847461e-05, | |
| "loss": 1.5928, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 0.0679310772282137, | |
| "grad_norm": 1.0705502033233643, | |
| "learning_rate": 2.476172311325783e-05, | |
| "loss": 1.6859, | |
| "step": 137 | |
| }, | |
| { | |
| "epoch": 0.06842692450725177, | |
| "grad_norm": 0.9703640341758728, | |
| "learning_rate": 2.405152131093926e-05, | |
| "loss": 1.5135, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 0.06892277178628982, | |
| "grad_norm": 1.0446672439575195, | |
| "learning_rate": 2.3348413563600325e-05, | |
| "loss": 1.5267, | |
| "step": 139 | |
| }, | |
| { | |
| "epoch": 0.06941861906532788, | |
| "grad_norm": 1.001079797744751, | |
| "learning_rate": 2.2652592093878666e-05, | |
| "loss": 1.4028, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.06991446634436593, | |
| "grad_norm": 1.0602779388427734, | |
| "learning_rate": 2.196424713241637e-05, | |
| "loss": 1.5749, | |
| "step": 141 | |
| }, | |
| { | |
| "epoch": 0.070410313623404, | |
| "grad_norm": 1.1383336782455444, | |
| "learning_rate": 2.128356686585282e-05, | |
| "loss": 1.5793, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 0.07090616090244205, | |
| "grad_norm": 1.0032070875167847, | |
| "learning_rate": 2.061073738537635e-05, | |
| "loss": 1.496, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 0.0714020081814801, | |
| "grad_norm": 1.1157033443450928, | |
| "learning_rate": 1.9945942635848748e-05, | |
| "loss": 1.6478, | |
| "step": 144 | |
| }, | |
| { | |
| "epoch": 0.07189785546051816, | |
| "grad_norm": 1.0024540424346924, | |
| "learning_rate": 1.928936436551661e-05, | |
| "loss": 1.5784, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 0.07239370273955621, | |
| "grad_norm": 1.0845361948013306, | |
| "learning_rate": 1.8641182076323148e-05, | |
| "loss": 1.6175, | |
| "step": 146 | |
| }, | |
| { | |
| "epoch": 0.07288955001859428, | |
| "grad_norm": 1.1753878593444824, | |
| "learning_rate": 1.800157297483417e-05, | |
| "loss": 1.7581, | |
| "step": 147 | |
| }, | |
| { | |
| "epoch": 0.07338539729763233, | |
| "grad_norm": 1.1944782733917236, | |
| "learning_rate": 1.7370711923791567e-05, | |
| "loss": 1.5709, | |
| "step": 148 | |
| }, | |
| { | |
| "epoch": 0.07388124457667039, | |
| "grad_norm": 1.0440027713775635, | |
| "learning_rate": 1.6748771394307585e-05, | |
| "loss": 1.6612, | |
| "step": 149 | |
| }, | |
| { | |
| "epoch": 0.07437709185570844, | |
| "grad_norm": 1.1899319887161255, | |
| "learning_rate": 1.6135921418712956e-05, | |
| "loss": 1.5008, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.07437709185570844, | |
| "eval_loss": 1.5300506353378296, | |
| "eval_runtime": 101.4061, | |
| "eval_samples_per_second": 33.499, | |
| "eval_steps_per_second": 8.382, | |
| "step": 150 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 5, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.102581815083008e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |