SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/google_t5
/t5_small_amazon
/trainer_state.json
| { | |
| "best_metric": 0.7013603448867798, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/google_t5/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": 2.4166171550750732, | |
| "learning_rate": 0.0004956140350877193, | |
| "loss": 3.1083, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.1570193767547607, | |
| "learning_rate": 0.0004912280701754386, | |
| "loss": 2.9816, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 1.6915223598480225, | |
| "learning_rate": 0.0004868421052631579, | |
| "loss": 2.9485, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 1.6796109676361084, | |
| "learning_rate": 0.0004824561403508772, | |
| "loss": 2.5652, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 2.0288569927215576, | |
| "learning_rate": 0.00047807017543859647, | |
| "loss": 2.1202, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "eval_accuracy": 0.5329380764163373, | |
| "eval_f1_macro": 0.384104867500989, | |
| "eval_f1_micro": 0.5329380764163373, | |
| "eval_loss": 1.808777093887329, | |
| "eval_runtime": 1.2462, | |
| "eval_samples_per_second": 1218.095, | |
| "eval_steps_per_second": 38.517, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 2.4368162155151367, | |
| "learning_rate": 0.00047368421052631577, | |
| "loss": 1.8285, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 2.4492692947387695, | |
| "learning_rate": 0.0004692982456140351, | |
| "loss": 1.5762, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 4.4139299392700195, | |
| "learning_rate": 0.00046491228070175437, | |
| "loss": 1.3461, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 2.3908071517944336, | |
| "learning_rate": 0.0004605263157894737, | |
| "loss": 1.3544, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 2.259780168533325, | |
| "learning_rate": 0.000456140350877193, | |
| "loss": 1.1993, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_accuracy": 0.6613965744400527, | |
| "eval_f1_macro": 0.5438670494720591, | |
| "eval_f1_micro": 0.6613965744400527, | |
| "eval_loss": 1.1745669841766357, | |
| "eval_runtime": 1.2956, | |
| "eval_samples_per_second": 1171.684, | |
| "eval_steps_per_second": 37.049, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 2.785118341445923, | |
| "learning_rate": 0.00045175438596491233, | |
| "loss": 1.278, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 3.2425591945648193, | |
| "learning_rate": 0.0004473684210526316, | |
| "loss": 1.0901, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 3.0471508502960205, | |
| "learning_rate": 0.0004429824561403509, | |
| "loss": 1.1888, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 2.5289652347564697, | |
| "learning_rate": 0.0004385964912280702, | |
| "loss": 0.9788, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 2.7128379344940186, | |
| "learning_rate": 0.0004342105263157895, | |
| "loss": 1.0698, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "eval_accuracy": 0.6884057971014492, | |
| "eval_f1_macro": 0.6003720421134128, | |
| "eval_f1_micro": 0.6884057971014492, | |
| "eval_loss": 1.0115997791290283, | |
| "eval_runtime": 1.2952, | |
| "eval_samples_per_second": 1172.016, | |
| "eval_steps_per_second": 37.06, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 3.8709490299224854, | |
| "learning_rate": 0.0004298245614035088, | |
| "loss": 1.0113, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 3.408216714859009, | |
| "learning_rate": 0.0004254385964912281, | |
| "loss": 1.0532, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 2.8890082836151123, | |
| "learning_rate": 0.00042105263157894734, | |
| "loss": 1.0167, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 4.663267612457275, | |
| "learning_rate": 0.0004166666666666667, | |
| "loss": 1.0561, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 2.8700387477874756, | |
| "learning_rate": 0.000412280701754386, | |
| "loss": 0.8999, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.717391304347826, | |
| "eval_f1_macro": 0.6538801770527956, | |
| "eval_f1_micro": 0.717391304347826, | |
| "eval_loss": 0.9428409337997437, | |
| "eval_runtime": 1.2446, | |
| "eval_samples_per_second": 1219.646, | |
| "eval_steps_per_second": 38.566, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 3.236957550048828, | |
| "learning_rate": 0.00040789473684210524, | |
| "loss": 0.8635, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 5.146123886108398, | |
| "learning_rate": 0.00040350877192982455, | |
| "loss": 1.0005, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 3.9536075592041016, | |
| "learning_rate": 0.0003991228070175439, | |
| "loss": 0.9122, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 2.807563304901123, | |
| "learning_rate": 0.00039473684210526315, | |
| "loss": 0.9735, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 7.1934919357299805, | |
| "learning_rate": 0.00039035087719298245, | |
| "loss": 1.1022, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "eval_accuracy": 0.7246376811594203, | |
| "eval_f1_macro": 0.6715689496561603, | |
| "eval_f1_micro": 0.7246376811594203, | |
| "eval_loss": 0.8931711316108704, | |
| "eval_runtime": 1.296, | |
| "eval_samples_per_second": 1171.291, | |
| "eval_steps_per_second": 37.037, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 2.2508842945098877, | |
| "learning_rate": 0.00038596491228070175, | |
| "loss": 0.9077, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 3.2862801551818848, | |
| "learning_rate": 0.00038157894736842105, | |
| "loss": 0.9932, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 2.539895534515381, | |
| "learning_rate": 0.00037719298245614036, | |
| "loss": 0.8718, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 4.113556385040283, | |
| "learning_rate": 0.00037280701754385966, | |
| "loss": 0.856, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 2.2022409439086914, | |
| "learning_rate": 0.00036842105263157896, | |
| "loss": 0.8337, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_accuracy": 0.7285902503293807, | |
| "eval_f1_macro": 0.6788777576959576, | |
| "eval_f1_micro": 0.7285902503293807, | |
| "eval_loss": 0.8664311766624451, | |
| "eval_runtime": 1.2586, | |
| "eval_samples_per_second": 1206.118, | |
| "eval_steps_per_second": 38.138, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 3.421111583709717, | |
| "learning_rate": 0.00036403508771929826, | |
| "loss": 0.836, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 4.356611251831055, | |
| "learning_rate": 0.00035964912280701756, | |
| "loss": 0.911, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 4.17406702041626, | |
| "learning_rate": 0.00035526315789473687, | |
| "loss": 0.9708, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 3.6954843997955322, | |
| "learning_rate": 0.0003508771929824561, | |
| "loss": 0.8783, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 3.14982533454895, | |
| "learning_rate": 0.00034649122807017547, | |
| "loss": 0.9594, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "eval_accuracy": 0.7503293807641633, | |
| "eval_f1_macro": 0.6993639076591665, | |
| "eval_f1_micro": 0.7503293807641633, | |
| "eval_loss": 0.8445017337799072, | |
| "eval_runtime": 1.2645, | |
| "eval_samples_per_second": 1200.522, | |
| "eval_steps_per_second": 37.961, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 3.1979424953460693, | |
| "learning_rate": 0.00034210526315789477, | |
| "loss": 0.7379, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.852599620819092, | |
| "learning_rate": 0.000337719298245614, | |
| "loss": 0.931, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.8022282123565674, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.7579, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "grad_norm": 3.0786988735198975, | |
| "learning_rate": 0.0003289473684210527, | |
| "loss": 0.7347, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 2.600757360458374, | |
| "learning_rate": 0.0003245614035087719, | |
| "loss": 0.803, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_accuracy": 0.7529644268774703, | |
| "eval_f1_macro": 0.6995912826384776, | |
| "eval_f1_micro": 0.7529644268774703, | |
| "eval_loss": 0.8047745227813721, | |
| "eval_runtime": 1.3204, | |
| "eval_samples_per_second": 1149.687, | |
| "eval_steps_per_second": 36.354, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 2.7453322410583496, | |
| "learning_rate": 0.00032017543859649123, | |
| "loss": 0.7006, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 2.7500882148742676, | |
| "learning_rate": 0.00031578947368421053, | |
| "loss": 0.7501, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "grad_norm": 4.403367519378662, | |
| "learning_rate": 0.00031140350877192983, | |
| "loss": 0.8383, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "grad_norm": 2.546445369720459, | |
| "learning_rate": 0.00030701754385964913, | |
| "loss": 0.6768, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 1.9148272275924683, | |
| "learning_rate": 0.00030263157894736844, | |
| "loss": 0.7271, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "eval_accuracy": 0.7602108036890646, | |
| "eval_f1_macro": 0.7019487651738747, | |
| "eval_f1_micro": 0.7602108036890646, | |
| "eval_loss": 0.7776466608047485, | |
| "eval_runtime": 1.3229, | |
| "eval_samples_per_second": 1147.482, | |
| "eval_steps_per_second": 36.284, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 3.360862970352173, | |
| "learning_rate": 0.0002982456140350877, | |
| "loss": 0.8197, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "grad_norm": 2.3219194412231445, | |
| "learning_rate": 0.00029385964912280704, | |
| "loss": 0.6231, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "grad_norm": 4.6816229820251465, | |
| "learning_rate": 0.00028947368421052634, | |
| "loss": 0.7613, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "grad_norm": 3.6067404747009277, | |
| "learning_rate": 0.00028508771929824564, | |
| "loss": 0.6973, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "grad_norm": 2.7148826122283936, | |
| "learning_rate": 0.0002807017543859649, | |
| "loss": 0.6694, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "eval_accuracy": 0.7608695652173914, | |
| "eval_f1_macro": 0.7083962744871815, | |
| "eval_f1_micro": 0.7608695652173914, | |
| "eval_loss": 0.7674228549003601, | |
| "eval_runtime": 1.272, | |
| "eval_samples_per_second": 1193.385, | |
| "eval_steps_per_second": 37.735, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 2.637134552001953, | |
| "learning_rate": 0.00027631578947368425, | |
| "loss": 0.6469, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 2.232243776321411, | |
| "learning_rate": 0.00027192982456140355, | |
| "loss": 0.7557, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 3.86612606048584, | |
| "learning_rate": 0.0002675438596491228, | |
| "loss": 0.7325, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "grad_norm": 2.2842824459075928, | |
| "learning_rate": 0.0002631578947368421, | |
| "loss": 0.5918, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 2.8408637046813965, | |
| "learning_rate": 0.00025877192982456146, | |
| "loss": 0.6109, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "eval_accuracy": 0.7608695652173914, | |
| "eval_f1_macro": 0.7081134075783231, | |
| "eval_f1_micro": 0.7608695652173914, | |
| "eval_loss": 0.7647947669029236, | |
| "eval_runtime": 1.2692, | |
| "eval_samples_per_second": 1196.015, | |
| "eval_steps_per_second": 37.819, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 2.67587947845459, | |
| "learning_rate": 0.0002543859649122807, | |
| "loss": 0.6937, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 2.2099223136901855, | |
| "learning_rate": 0.00025, | |
| "loss": 0.6696, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 2.383314609527588, | |
| "learning_rate": 0.0002456140350877193, | |
| "loss": 0.6252, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 3.261552572250366, | |
| "learning_rate": 0.0002412280701754386, | |
| "loss": 0.5798, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 2.702723741531372, | |
| "learning_rate": 0.00023684210526315788, | |
| "loss": 0.6575, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_accuracy": 0.7628458498023716, | |
| "eval_f1_macro": 0.71173444087918, | |
| "eval_f1_micro": 0.7628458498023716, | |
| "eval_loss": 0.7526513934135437, | |
| "eval_runtime": 1.4987, | |
| "eval_samples_per_second": 1012.867, | |
| "eval_steps_per_second": 32.027, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "grad_norm": 2.8729443550109863, | |
| "learning_rate": 0.00023245614035087719, | |
| "loss": 0.6817, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 2.480459213256836, | |
| "learning_rate": 0.0002280701754385965, | |
| "loss": 0.6792, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "grad_norm": 3.253549575805664, | |
| "learning_rate": 0.0002236842105263158, | |
| "loss": 0.6368, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 3.7337493896484375, | |
| "learning_rate": 0.0002192982456140351, | |
| "loss": 0.7799, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 3.4292104244232178, | |
| "learning_rate": 0.0002149122807017544, | |
| "loss": 0.777, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "eval_accuracy": 0.769433465085639, | |
| "eval_f1_macro": 0.7217630318046797, | |
| "eval_f1_micro": 0.769433465085639, | |
| "eval_loss": 0.7418988943099976, | |
| "eval_runtime": 1.2675, | |
| "eval_samples_per_second": 1197.616, | |
| "eval_steps_per_second": 37.869, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "grad_norm": 2.52197003364563, | |
| "learning_rate": 0.00021052631578947367, | |
| "loss": 0.5507, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "grad_norm": 3.5961380004882812, | |
| "learning_rate": 0.000206140350877193, | |
| "loss": 0.6555, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 5.47299337387085, | |
| "learning_rate": 0.00020175438596491227, | |
| "loss": 0.748, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 3.6607069969177246, | |
| "learning_rate": 0.00019736842105263157, | |
| "loss": 0.6491, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 3.5043647289276123, | |
| "learning_rate": 0.00019298245614035088, | |
| "loss": 0.6362, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "eval_accuracy": 0.7799736495388669, | |
| "eval_f1_macro": 0.7301450284031649, | |
| "eval_f1_micro": 0.7799736495388669, | |
| "eval_loss": 0.7271929383277893, | |
| "eval_runtime": 1.2681, | |
| "eval_samples_per_second": 1197.064, | |
| "eval_steps_per_second": 37.852, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "grad_norm": 2.762425661087036, | |
| "learning_rate": 0.00018859649122807018, | |
| "loss": 0.7247, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 2.687269926071167, | |
| "learning_rate": 0.00018421052631578948, | |
| "loss": 0.6264, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 2.9003891944885254, | |
| "learning_rate": 0.00017982456140350878, | |
| "loss": 0.6243, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 3.443100929260254, | |
| "learning_rate": 0.00017543859649122806, | |
| "loss": 0.7026, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 2.562163829803467, | |
| "learning_rate": 0.00017105263157894739, | |
| "loss": 0.648, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "eval_accuracy": 0.7812911725955204, | |
| "eval_f1_macro": 0.7356498242584253, | |
| "eval_f1_micro": 0.7812911725955204, | |
| "eval_loss": 0.7136700749397278, | |
| "eval_runtime": 1.268, | |
| "eval_samples_per_second": 1197.205, | |
| "eval_steps_per_second": 37.856, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 2.538322925567627, | |
| "learning_rate": 0.00016666666666666666, | |
| "loss": 0.6682, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.03, | |
| "grad_norm": 2.9608592987060547, | |
| "learning_rate": 0.00016228070175438596, | |
| "loss": 0.554, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 3.1620254516601562, | |
| "learning_rate": 0.00015789473684210527, | |
| "loss": 0.5446, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 2.561002016067505, | |
| "learning_rate": 0.00015350877192982457, | |
| "loss": 0.5153, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 2.196606397628784, | |
| "learning_rate": 0.00014912280701754384, | |
| "loss": 0.4981, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "eval_accuracy": 0.7766798418972332, | |
| "eval_f1_macro": 0.7258055584715567, | |
| "eval_f1_micro": 0.7766798418972332, | |
| "eval_loss": 0.7154478430747986, | |
| "eval_runtime": 1.3162, | |
| "eval_samples_per_second": 1153.362, | |
| "eval_steps_per_second": 36.47, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "grad_norm": 2.3084490299224854, | |
| "learning_rate": 0.00014473684210526317, | |
| "loss": 0.4972, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 2.283141613006592, | |
| "learning_rate": 0.00014035087719298245, | |
| "loss": 0.4606, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "grad_norm": 4.426635265350342, | |
| "learning_rate": 0.00013596491228070177, | |
| "loss": 0.5108, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 3.8946399688720703, | |
| "learning_rate": 0.00013157894736842105, | |
| "loss": 0.5072, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 2.5790274143218994, | |
| "learning_rate": 0.00012719298245614035, | |
| "loss": 0.4955, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_accuracy": 0.7799736495388669, | |
| "eval_f1_macro": 0.731846151928883, | |
| "eval_f1_micro": 0.7799736495388669, | |
| "eval_loss": 0.7233121395111084, | |
| "eval_runtime": 1.2667, | |
| "eval_samples_per_second": 1198.367, | |
| "eval_steps_per_second": 37.893, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 3.8150198459625244, | |
| "learning_rate": 0.00012280701754385965, | |
| "loss": 0.4035, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "grad_norm": 2.8765792846679688, | |
| "learning_rate": 0.00011842105263157894, | |
| "loss": 0.5247, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 3.2408721446990967, | |
| "learning_rate": 0.00011403508771929824, | |
| "loss": 0.5062, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "grad_norm": 2.066289186477661, | |
| "learning_rate": 0.00010964912280701755, | |
| "loss": 0.3746, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 4.370830535888672, | |
| "learning_rate": 0.00010526315789473683, | |
| "loss": 0.4451, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_accuracy": 0.7779973649538867, | |
| "eval_f1_macro": 0.7279848994870372, | |
| "eval_f1_micro": 0.7779973649538867, | |
| "eval_loss": 0.7181668877601624, | |
| "eval_runtime": 1.3175, | |
| "eval_samples_per_second": 1152.141, | |
| "eval_steps_per_second": 36.431, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "grad_norm": 2.3764281272888184, | |
| "learning_rate": 0.00010087719298245614, | |
| "loss": 0.5377, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 3.0116217136383057, | |
| "learning_rate": 9.649122807017544e-05, | |
| "loss": 0.5044, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 2.8177366256713867, | |
| "learning_rate": 9.210526315789474e-05, | |
| "loss": 0.5144, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "grad_norm": 3.101243257522583, | |
| "learning_rate": 8.771929824561403e-05, | |
| "loss": 0.6122, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 2.797656297683716, | |
| "learning_rate": 8.333333333333333e-05, | |
| "loss": 0.421, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_accuracy": 0.7747035573122529, | |
| "eval_f1_macro": 0.7261997754426944, | |
| "eval_f1_micro": 0.7747035573122529, | |
| "eval_loss": 0.7116832733154297, | |
| "eval_runtime": 1.2665, | |
| "eval_samples_per_second": 1198.612, | |
| "eval_steps_per_second": 37.901, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 1.7224795818328857, | |
| "learning_rate": 7.894736842105263e-05, | |
| "loss": 0.5545, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 1.7975401878356934, | |
| "learning_rate": 7.456140350877192e-05, | |
| "loss": 0.5856, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 3.113168478012085, | |
| "learning_rate": 7.017543859649122e-05, | |
| "loss": 0.4921, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 2.885319709777832, | |
| "learning_rate": 6.578947368421052e-05, | |
| "loss": 0.4638, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "grad_norm": 3.5493428707122803, | |
| "learning_rate": 6.140350877192983e-05, | |
| "loss": 0.4853, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "eval_accuracy": 0.7760210803689065, | |
| "eval_f1_macro": 0.7271826503982889, | |
| "eval_f1_micro": 0.7760210803689065, | |
| "eval_loss": 0.7091566324234009, | |
| "eval_runtime": 1.3164, | |
| "eval_samples_per_second": 1153.153, | |
| "eval_steps_per_second": 36.463, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.66, | |
| "grad_norm": 3.597778081893921, | |
| "learning_rate": 5.701754385964912e-05, | |
| "loss": 0.4753, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "grad_norm": 2.5180883407592773, | |
| "learning_rate": 5.263157894736842e-05, | |
| "loss": 0.5155, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "grad_norm": 2.6854727268218994, | |
| "learning_rate": 4.824561403508772e-05, | |
| "loss": 0.4774, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 3.311980962753296, | |
| "learning_rate": 4.3859649122807014e-05, | |
| "loss": 0.524, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "grad_norm": 4.007936477661133, | |
| "learning_rate": 3.9473684210526316e-05, | |
| "loss": 0.5442, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "eval_accuracy": 0.7740447957839263, | |
| "eval_f1_macro": 0.7272104730436151, | |
| "eval_f1_micro": 0.7740447957839263, | |
| "eval_loss": 0.7113538384437561, | |
| "eval_runtime": 1.3148, | |
| "eval_samples_per_second": 1154.575, | |
| "eval_steps_per_second": 36.508, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "grad_norm": 2.6298959255218506, | |
| "learning_rate": 3.508771929824561e-05, | |
| "loss": 0.5173, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 2.9783828258514404, | |
| "learning_rate": 3.0701754385964913e-05, | |
| "loss": 0.6384, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "grad_norm": 3.03838849067688, | |
| "learning_rate": 2.631578947368421e-05, | |
| "loss": 0.5627, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "grad_norm": 2.553374767303467, | |
| "learning_rate": 2.1929824561403507e-05, | |
| "loss": 0.4947, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "grad_norm": 2.0844762325286865, | |
| "learning_rate": 1.7543859649122806e-05, | |
| "loss": 0.4863, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "eval_accuracy": 0.7766798418972332, | |
| "eval_f1_macro": 0.7273078975523218, | |
| "eval_f1_micro": 0.7766798418972332, | |
| "eval_loss": 0.7013603448867798, | |
| "eval_runtime": 1.3157, | |
| "eval_samples_per_second": 1153.75, | |
| "eval_steps_per_second": 36.482, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 1.7584428787231445, | |
| "learning_rate": 1.3157894736842104e-05, | |
| "loss": 0.516, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 2.012805700302124, | |
| "learning_rate": 8.771929824561403e-06, | |
| "loss": 0.6011, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 2.97, | |
| "grad_norm": 3.4424493312835693, | |
| "learning_rate": 4.3859649122807014e-06, | |
| "loss": 0.4124, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.8425490856170654, | |
| "learning_rate": 0.0, | |
| "loss": 0.4705, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1140, | |
| "total_flos": 1242006489661440.0, | |
| "train_loss": 0.8189868109268055, | |
| "train_runtime": 138.2392, | |
| "train_samples_per_second": 263.543, | |
| "train_steps_per_second": 8.247 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1140, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 1242006489661440.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |