| { | |
| "best_metric": 0.6518497467041016, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google/flan_t5_small_amazon/checkpoint-1100", | |
| "epoch": 3.0, | |
| "eval_steps": 50, | |
| "global_step": 1140, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 3.3804776668548584, | |
| "learning_rate": 0.0004956140350877193, | |
| "loss": 3.1062, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.343827962875366, | |
| "learning_rate": 0.0004912280701754386, | |
| "loss": 2.9891, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 1.987942099571228, | |
| "learning_rate": 0.0004868421052631579, | |
| "loss": 2.8938, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 2.4765126705169678, | |
| "learning_rate": 0.0004824561403508772, | |
| "loss": 2.4308, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 3.325758934020996, | |
| "learning_rate": 0.00047807017543859647, | |
| "loss": 1.9282, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "eval_accuracy": 0.5737812911725956, | |
| "eval_f1_macro": 0.44362854909805843, | |
| "eval_f1_micro": 0.5737812911725956, | |
| "eval_loss": 1.590394139289856, | |
| "eval_runtime": 1.5287, | |
| "eval_samples_per_second": 992.985, | |
| "eval_steps_per_second": 31.399, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 3.0366220474243164, | |
| "learning_rate": 0.00047368421052631577, | |
| "loss": 1.6028, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 3.9418632984161377, | |
| "learning_rate": 0.0004692982456140351, | |
| "loss": 1.3737, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 3.228276014328003, | |
| "learning_rate": 0.00046491228070175437, | |
| "loss": 1.2147, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 3.007902145385742, | |
| "learning_rate": 0.0004605263157894737, | |
| "loss": 1.2373, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 3.132214307785034, | |
| "learning_rate": 0.000456140350877193, | |
| "loss": 1.1323, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_accuracy": 0.663372859025033, | |
| "eval_f1_macro": 0.5640942564973862, | |
| "eval_f1_micro": 0.663372859025033, | |
| "eval_loss": 1.1076021194458008, | |
| "eval_runtime": 1.5766, | |
| "eval_samples_per_second": 962.833, | |
| "eval_steps_per_second": 30.445, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 4.991384506225586, | |
| "learning_rate": 0.00045175438596491233, | |
| "loss": 1.21, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 4.180238723754883, | |
| "learning_rate": 0.0004473684210526316, | |
| "loss": 1.0477, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 3.497032642364502, | |
| "learning_rate": 0.0004429824561403509, | |
| "loss": 1.1554, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 3.396751642227173, | |
| "learning_rate": 0.0004385964912280702, | |
| "loss": 0.9906, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 2.457590341567993, | |
| "learning_rate": 0.0004342105263157895, | |
| "loss": 0.9976, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "eval_accuracy": 0.7206851119894598, | |
| "eval_f1_macro": 0.646835154821118, | |
| "eval_f1_micro": 0.7206851119894598, | |
| "eval_loss": 0.9464844465255737, | |
| "eval_runtime": 1.5277, | |
| "eval_samples_per_second": 993.644, | |
| "eval_steps_per_second": 31.42, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 4.1680402755737305, | |
| "learning_rate": 0.0004298245614035088, | |
| "loss": 0.9004, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 3.452507734298706, | |
| "learning_rate": 0.0004254385964912281, | |
| "loss": 1.0064, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 2.8887767791748047, | |
| "learning_rate": 0.00042105263157894734, | |
| "loss": 0.9662, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 5.135780334472656, | |
| "learning_rate": 0.0004166666666666667, | |
| "loss": 1.0267, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 3.5169897079467773, | |
| "learning_rate": 0.000412280701754386, | |
| "loss": 0.928, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.733201581027668, | |
| "eval_f1_macro": 0.6859595929599217, | |
| "eval_f1_micro": 0.733201581027668, | |
| "eval_loss": 0.8840087652206421, | |
| "eval_runtime": 1.5811, | |
| "eval_samples_per_second": 960.099, | |
| "eval_steps_per_second": 30.359, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 3.8375213146209717, | |
| "learning_rate": 0.00040789473684210524, | |
| "loss": 0.8223, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 3.949693441390991, | |
| "learning_rate": 0.00040350877192982455, | |
| "loss": 0.9726, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 3.1758928298950195, | |
| "learning_rate": 0.0003991228070175439, | |
| "loss": 0.8881, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 4.932584285736084, | |
| "learning_rate": 0.00039473684210526315, | |
| "loss": 0.9233, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 3.125002861022949, | |
| "learning_rate": 0.00039035087719298245, | |
| "loss": 0.974, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "eval_accuracy": 0.7523056653491436, | |
| "eval_f1_macro": 0.7025921892520358, | |
| "eval_f1_micro": 0.7523056653491436, | |
| "eval_loss": 0.8179047107696533, | |
| "eval_runtime": 1.5845, | |
| "eval_samples_per_second": 958.002, | |
| "eval_steps_per_second": 30.293, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 3.811002016067505, | |
| "learning_rate": 0.00038596491228070175, | |
| "loss": 0.8885, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 2.872647762298584, | |
| "learning_rate": 0.00038157894736842105, | |
| "loss": 0.9161, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 3.938746452331543, | |
| "learning_rate": 0.00037719298245614036, | |
| "loss": 0.7768, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 3.9342243671417236, | |
| "learning_rate": 0.00037280701754385966, | |
| "loss": 0.8114, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 3.424884557723999, | |
| "learning_rate": 0.00036842105263157896, | |
| "loss": 0.8206, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_accuracy": 0.7674571805006588, | |
| "eval_f1_macro": 0.7181951416155129, | |
| "eval_f1_micro": 0.7674571805006588, | |
| "eval_loss": 0.7982419729232788, | |
| "eval_runtime": 1.5868, | |
| "eval_samples_per_second": 956.662, | |
| "eval_steps_per_second": 30.25, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 4.263403415679932, | |
| "learning_rate": 0.00036403508771929826, | |
| "loss": 0.7327, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 3.17680025100708, | |
| "learning_rate": 0.00035964912280701756, | |
| "loss": 0.8797, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 3.7559242248535156, | |
| "learning_rate": 0.00035526315789473687, | |
| "loss": 0.9148, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 3.6917154788970947, | |
| "learning_rate": 0.0003508771929824561, | |
| "loss": 0.7763, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 3.3386855125427246, | |
| "learning_rate": 0.00034649122807017547, | |
| "loss": 0.8863, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "eval_accuracy": 0.7773386034255599, | |
| "eval_f1_macro": 0.7301140464284491, | |
| "eval_f1_micro": 0.7773386034255599, | |
| "eval_loss": 0.7444999814033508, | |
| "eval_runtime": 1.593, | |
| "eval_samples_per_second": 952.941, | |
| "eval_steps_per_second": 30.133, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 2.625443935394287, | |
| "learning_rate": 0.00034210526315789477, | |
| "loss": 0.7353, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 4.095985412597656, | |
| "learning_rate": 0.000337719298245614, | |
| "loss": 0.9274, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.8155393600463867, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.7887, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "grad_norm": 3.3496756553649902, | |
| "learning_rate": 0.0003289473684210527, | |
| "loss": 0.682, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 3.9287924766540527, | |
| "learning_rate": 0.0003245614035087719, | |
| "loss": 0.713, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_accuracy": 0.7740447957839263, | |
| "eval_f1_macro": 0.7390551518744245, | |
| "eval_f1_micro": 0.7740447957839263, | |
| "eval_loss": 0.7427995204925537, | |
| "eval_runtime": 1.5483, | |
| "eval_samples_per_second": 980.422, | |
| "eval_steps_per_second": 31.001, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 3.985713243484497, | |
| "learning_rate": 0.00032017543859649123, | |
| "loss": 0.597, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 3.647797107696533, | |
| "learning_rate": 0.00031578947368421053, | |
| "loss": 0.6752, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 3.2907509803771973, | |
| "learning_rate": 0.00031140350877192983, | |
| "loss": 0.6903, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "grad_norm": 4.062851905822754, | |
| "learning_rate": 0.00030701754385964913, | |
| "loss": 0.5708, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 3.1373462677001953, | |
| "learning_rate": 0.00030263157894736844, | |
| "loss": 0.6544, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "eval_accuracy": 0.7852437417654808, | |
| "eval_f1_macro": 0.7379347673603733, | |
| "eval_f1_micro": 0.7852437417654808, | |
| "eval_loss": 0.7234027981758118, | |
| "eval_runtime": 1.5405, | |
| "eval_samples_per_second": 985.365, | |
| "eval_steps_per_second": 31.158, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 3.7265396118164062, | |
| "learning_rate": 0.0002982456140350877, | |
| "loss": 0.7901, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "grad_norm": 2.8776745796203613, | |
| "learning_rate": 0.00029385964912280704, | |
| "loss": 0.5103, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "grad_norm": 4.379913806915283, | |
| "learning_rate": 0.00028947368421052634, | |
| "loss": 0.7062, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 3.8924524784088135, | |
| "learning_rate": 0.00028508771929824564, | |
| "loss": 0.657, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "grad_norm": 4.120116233825684, | |
| "learning_rate": 0.0002807017543859649, | |
| "loss": 0.6034, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "eval_accuracy": 0.7924901185770751, | |
| "eval_f1_macro": 0.7648022166570957, | |
| "eval_f1_micro": 0.7924901185770751, | |
| "eval_loss": 0.7140358090400696, | |
| "eval_runtime": 1.5497, | |
| "eval_samples_per_second": 979.546, | |
| "eval_steps_per_second": 30.974, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 3.7763350009918213, | |
| "learning_rate": 0.00027631578947368425, | |
| "loss": 0.5917, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 3.2136292457580566, | |
| "learning_rate": 0.00027192982456140355, | |
| "loss": 0.7325, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 5.103039264678955, | |
| "learning_rate": 0.0002675438596491228, | |
| "loss": 0.7051, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "grad_norm": 2.6361660957336426, | |
| "learning_rate": 0.0002631578947368421, | |
| "loss": 0.5079, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 3.617443561553955, | |
| "learning_rate": 0.00025877192982456146, | |
| "loss": 0.588, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "eval_accuracy": 0.7931488801054019, | |
| "eval_f1_macro": 0.7585136991271696, | |
| "eval_f1_micro": 0.7931488801054019, | |
| "eval_loss": 0.7061514258384705, | |
| "eval_runtime": 1.5944, | |
| "eval_samples_per_second": 952.063, | |
| "eval_steps_per_second": 30.105, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 3.122095823287964, | |
| "learning_rate": 0.0002543859649122807, | |
| "loss": 0.6284, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 2.7279186248779297, | |
| "learning_rate": 0.00025, | |
| "loss": 0.6255, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 3.1004347801208496, | |
| "learning_rate": 0.0002456140350877193, | |
| "loss": 0.5774, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 3.556870937347412, | |
| "learning_rate": 0.0002412280701754386, | |
| "loss": 0.5078, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 4.2002854347229, | |
| "learning_rate": 0.00023684210526315788, | |
| "loss": 0.6035, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_accuracy": 0.7924901185770751, | |
| "eval_f1_macro": 0.7479706258386674, | |
| "eval_f1_micro": 0.7924901185770751, | |
| "eval_loss": 0.711154043674469, | |
| "eval_runtime": 1.6011, | |
| "eval_samples_per_second": 948.117, | |
| "eval_steps_per_second": 29.98, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 3.03747296333313, | |
| "learning_rate": 0.00023245614035087719, | |
| "loss": 0.6233, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 2.1721749305725098, | |
| "learning_rate": 0.0002280701754385965, | |
| "loss": 0.6446, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "grad_norm": 2.81701922416687, | |
| "learning_rate": 0.0002236842105263158, | |
| "loss": 0.6268, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 4.8511128425598145, | |
| "learning_rate": 0.0002192982456140351, | |
| "loss": 0.6635, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 3.052924871444702, | |
| "learning_rate": 0.0002149122807017544, | |
| "loss": 0.6616, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_accuracy": 0.7938076416337286, | |
| "eval_f1_macro": 0.7577530795114652, | |
| "eval_f1_micro": 0.7938076416337286, | |
| "eval_loss": 0.6783204674720764, | |
| "eval_runtime": 1.6022, | |
| "eval_samples_per_second": 947.452, | |
| "eval_steps_per_second": 29.959, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "grad_norm": 3.520653247833252, | |
| "learning_rate": 0.00021052631578947367, | |
| "loss": 0.5442, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 2.6852402687072754, | |
| "learning_rate": 0.000206140350877193, | |
| "loss": 0.5909, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 2.394245147705078, | |
| "learning_rate": 0.00020175438596491227, | |
| "loss": 0.6419, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 3.5546963214874268, | |
| "learning_rate": 0.00019736842105263157, | |
| "loss": 0.6128, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 2.6713662147521973, | |
| "learning_rate": 0.00019298245614035088, | |
| "loss": 0.6334, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "eval_accuracy": 0.8003952569169961, | |
| "eval_f1_macro": 0.7850643712314976, | |
| "eval_f1_micro": 0.8003952569169961, | |
| "eval_loss": 0.6815611720085144, | |
| "eval_runtime": 1.601, | |
| "eval_samples_per_second": 948.157, | |
| "eval_steps_per_second": 29.981, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 2.6882362365722656, | |
| "learning_rate": 0.00018859649122807018, | |
| "loss": 0.5586, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 3.613619565963745, | |
| "learning_rate": 0.00018421052631578948, | |
| "loss": 0.5914, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 2.6678318977355957, | |
| "learning_rate": 0.00017982456140350878, | |
| "loss": 0.6132, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 2.920417070388794, | |
| "learning_rate": 0.00017543859649122806, | |
| "loss": 0.6667, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 4.305054664611816, | |
| "learning_rate": 0.00017105263157894739, | |
| "loss": 0.5872, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "eval_accuracy": 0.8036890645586298, | |
| "eval_f1_macro": 0.7792442711200399, | |
| "eval_f1_micro": 0.8036890645586298, | |
| "eval_loss": 0.6531891822814941, | |
| "eval_runtime": 1.5978, | |
| "eval_samples_per_second": 950.034, | |
| "eval_steps_per_second": 30.041, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 2.5539097785949707, | |
| "learning_rate": 0.00016666666666666666, | |
| "loss": 0.5254, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "grad_norm": 3.1510868072509766, | |
| "learning_rate": 0.00016228070175438596, | |
| "loss": 0.4091, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 3.702329635620117, | |
| "learning_rate": 0.00015789473684210527, | |
| "loss": 0.4975, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 5.3907060623168945, | |
| "learning_rate": 0.00015350877192982457, | |
| "loss": 0.437, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 3.821787118911743, | |
| "learning_rate": 0.00014912280701754384, | |
| "loss": 0.4134, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "eval_accuracy": 0.8069828722002635, | |
| "eval_f1_macro": 0.7857989316065254, | |
| "eval_f1_micro": 0.8069828722002635, | |
| "eval_loss": 0.6601086258888245, | |
| "eval_runtime": 1.5931, | |
| "eval_samples_per_second": 952.843, | |
| "eval_steps_per_second": 30.129, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "grad_norm": 2.2001376152038574, | |
| "learning_rate": 0.00014473684210526317, | |
| "loss": 0.3875, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 2.7944624423980713, | |
| "learning_rate": 0.00014035087719298245, | |
| "loss": 0.4485, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 1.9566526412963867, | |
| "learning_rate": 0.00013596491228070177, | |
| "loss": 0.396, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 2.5819900035858154, | |
| "learning_rate": 0.00013157894736842105, | |
| "loss": 0.4155, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 3.8049263954162598, | |
| "learning_rate": 0.00012719298245614035, | |
| "loss": 0.518, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_accuracy": 0.8069828722002635, | |
| "eval_f1_macro": 0.7857680865770484, | |
| "eval_f1_micro": 0.8069828722002635, | |
| "eval_loss": 0.6772085428237915, | |
| "eval_runtime": 1.5894, | |
| "eval_samples_per_second": 955.055, | |
| "eval_steps_per_second": 30.199, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 4.169273853302002, | |
| "learning_rate": 0.00012280701754385965, | |
| "loss": 0.3547, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 3.642315626144409, | |
| "learning_rate": 0.00011842105263157894, | |
| "loss": 0.4423, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 3.0137033462524414, | |
| "learning_rate": 0.00011403508771929824, | |
| "loss": 0.4462, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 1.681169867515564, | |
| "learning_rate": 0.00010964912280701755, | |
| "loss": 0.3536, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 3.0964701175689697, | |
| "learning_rate": 0.00010526315789473683, | |
| "loss": 0.3891, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_accuracy": 0.8089591567852438, | |
| "eval_f1_macro": 0.7866273295700273, | |
| "eval_f1_micro": 0.8089591567852438, | |
| "eval_loss": 0.6751896142959595, | |
| "eval_runtime": 1.5446, | |
| "eval_samples_per_second": 982.806, | |
| "eval_steps_per_second": 31.077, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "grad_norm": 2.752586841583252, | |
| "learning_rate": 0.00010087719298245614, | |
| "loss": 0.4731, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 2.796949863433838, | |
| "learning_rate": 9.649122807017544e-05, | |
| "loss": 0.5025, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 3.236764430999756, | |
| "learning_rate": 9.210526315789474e-05, | |
| "loss": 0.4629, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "grad_norm": 3.6876580715179443, | |
| "learning_rate": 8.771929824561403e-05, | |
| "loss": 0.4943, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 4.9102911949157715, | |
| "learning_rate": 8.333333333333333e-05, | |
| "loss": 0.3389, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_accuracy": 0.8122529644268774, | |
| "eval_f1_macro": 0.7914121636262914, | |
| "eval_f1_micro": 0.8122529644268774, | |
| "eval_loss": 0.6639086008071899, | |
| "eval_runtime": 1.5986, | |
| "eval_samples_per_second": 949.6, | |
| "eval_steps_per_second": 30.027, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 1.5189881324768066, | |
| "learning_rate": 7.894736842105263e-05, | |
| "loss": 0.4331, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 3.3919639587402344, | |
| "learning_rate": 7.456140350877192e-05, | |
| "loss": 0.5003, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 4.042901992797852, | |
| "learning_rate": 7.017543859649122e-05, | |
| "loss": 0.4475, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 3.0205342769622803, | |
| "learning_rate": 6.578947368421052e-05, | |
| "loss": 0.4484, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "grad_norm": 3.377220630645752, | |
| "learning_rate": 6.140350877192983e-05, | |
| "loss": 0.4166, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "eval_accuracy": 0.8168642951251647, | |
| "eval_f1_macro": 0.8010465581589675, | |
| "eval_f1_micro": 0.8168642951251647, | |
| "eval_loss": 0.6590448021888733, | |
| "eval_runtime": 1.5979, | |
| "eval_samples_per_second": 950.026, | |
| "eval_steps_per_second": 30.04, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 3.584700584411621, | |
| "learning_rate": 5.701754385964912e-05, | |
| "loss": 0.3845, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "grad_norm": 2.4608445167541504, | |
| "learning_rate": 5.263157894736842e-05, | |
| "loss": 0.4711, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "grad_norm": 3.6688997745513916, | |
| "learning_rate": 4.824561403508772e-05, | |
| "loss": 0.4564, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 2.6104586124420166, | |
| "learning_rate": 4.3859649122807014e-05, | |
| "loss": 0.4098, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 4.012241363525391, | |
| "learning_rate": 3.9473684210526316e-05, | |
| "loss": 0.483, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "eval_accuracy": 0.8148880105401844, | |
| "eval_f1_macro": 0.7936908234810874, | |
| "eval_f1_micro": 0.8148880105401844, | |
| "eval_loss": 0.6629713177680969, | |
| "eval_runtime": 1.5955, | |
| "eval_samples_per_second": 951.432, | |
| "eval_steps_per_second": 30.085, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "grad_norm": 3.5108678340911865, | |
| "learning_rate": 3.508771929824561e-05, | |
| "loss": 0.4483, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 3.1787357330322266, | |
| "learning_rate": 3.0701754385964913e-05, | |
| "loss": 0.5206, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "grad_norm": 3.5252914428710938, | |
| "learning_rate": 2.631578947368421e-05, | |
| "loss": 0.5051, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 2.644866704940796, | |
| "learning_rate": 2.1929824561403507e-05, | |
| "loss": 0.4477, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "grad_norm": 3.4592483043670654, | |
| "learning_rate": 1.7543859649122806e-05, | |
| "loss": 0.4582, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "eval_accuracy": 0.8096179183135704, | |
| "eval_f1_macro": 0.788232456216749, | |
| "eval_f1_micro": 0.8096179183135704, | |
| "eval_loss": 0.6518497467041016, | |
| "eval_runtime": 1.537, | |
| "eval_samples_per_second": 987.659, | |
| "eval_steps_per_second": 31.23, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 1.930580973625183, | |
| "learning_rate": 1.3157894736842104e-05, | |
| "loss": 0.4766, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 3.2953085899353027, | |
| "learning_rate": 8.771929824561403e-06, | |
| "loss": 0.4975, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 2.97, | |
| "grad_norm": 3.153869390487671, | |
| "learning_rate": 4.3859649122807014e-06, | |
| "loss": 0.3219, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.2529447078704834, | |
| "learning_rate": 0.0, | |
| "loss": 0.3759, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1140, | |
| "total_flos": 1242149934858240.0, | |
| "train_loss": 0.7540260691391795, | |
| "train_runtime": 163.74, | |
| "train_samples_per_second": 222.499, | |
| "train_steps_per_second": 6.962 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1140, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 1242149934858240.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |