| { | |
| "best_metric": 0.777472734451294, | |
| "best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/distilbert/distilroberta_base_amazon/checkpoint-550", | |
| "epoch": 3.0, | |
| "eval_steps": 50, | |
| "global_step": 570, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.0744564533233643, | |
| "learning_rate": 1.9649122807017544e-05, | |
| "loss": 3.0919, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 3.013991355895996, | |
| "learning_rate": 1.929824561403509e-05, | |
| "loss": 3.0612, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 3.487683057785034, | |
| "learning_rate": 1.894736842105263e-05, | |
| "loss": 2.8423, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 3.85282564163208, | |
| "learning_rate": 1.8596491228070176e-05, | |
| "loss": 2.4944, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 3.4586243629455566, | |
| "learning_rate": 1.824561403508772e-05, | |
| "loss": 2.2364, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_accuracy": 0.5487483530961792, | |
| "eval_f1_macro": 0.3934336772383333, | |
| "eval_f1_micro": 0.5487483530961792, | |
| "eval_loss": 1.9915441274642944, | |
| "eval_runtime": 0.7632, | |
| "eval_samples_per_second": 1988.946, | |
| "eval_steps_per_second": 31.446, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 4.306815147399902, | |
| "learning_rate": 1.7894736842105264e-05, | |
| "loss": 1.9845, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 6.179586410522461, | |
| "learning_rate": 1.754385964912281e-05, | |
| "loss": 1.7504, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 4.895262718200684, | |
| "learning_rate": 1.719298245614035e-05, | |
| "loss": 1.5823, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 5.523647785186768, | |
| "learning_rate": 1.6842105263157896e-05, | |
| "loss": 1.5412, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 4.391608238220215, | |
| "learning_rate": 1.649122807017544e-05, | |
| "loss": 1.4061, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "eval_accuracy": 0.644927536231884, | |
| "eval_f1_macro": 0.5029573881708336, | |
| "eval_f1_micro": 0.644927536231884, | |
| "eval_loss": 1.3135813474655151, | |
| "eval_runtime": 0.812, | |
| "eval_samples_per_second": 1869.532, | |
| "eval_steps_per_second": 29.558, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 5.148378849029541, | |
| "learning_rate": 1.6140350877192984e-05, | |
| "loss": 1.3258, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 5.900285243988037, | |
| "learning_rate": 1.578947368421053e-05, | |
| "loss": 1.284, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 7.00700569152832, | |
| "learning_rate": 1.543859649122807e-05, | |
| "loss": 1.2605, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 5.3574113845825195, | |
| "learning_rate": 1.5087719298245615e-05, | |
| "loss": 1.2144, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 7.157119274139404, | |
| "learning_rate": 1.4736842105263159e-05, | |
| "loss": 1.1502, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_accuracy": 0.6936758893280632, | |
| "eval_f1_macro": 0.5696872174478277, | |
| "eval_f1_micro": 0.6936758893280632, | |
| "eval_loss": 1.1137052774429321, | |
| "eval_runtime": 0.7626, | |
| "eval_samples_per_second": 1990.548, | |
| "eval_steps_per_second": 31.471, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 7.763728618621826, | |
| "learning_rate": 1.4385964912280704e-05, | |
| "loss": 1.1604, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 7.102268218994141, | |
| "learning_rate": 1.4035087719298246e-05, | |
| "loss": 1.1256, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 7.185972690582275, | |
| "learning_rate": 1.3684210526315791e-05, | |
| "loss": 1.1088, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 8.286725997924805, | |
| "learning_rate": 1.3333333333333333e-05, | |
| "loss": 1.1016, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 6.963580131530762, | |
| "learning_rate": 1.2982456140350879e-05, | |
| "loss": 1.0396, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_accuracy": 0.735836627140975, | |
| "eval_f1_macro": 0.634528837939176, | |
| "eval_f1_micro": 0.735836627140975, | |
| "eval_loss": 0.9887633919715881, | |
| "eval_runtime": 0.7647, | |
| "eval_samples_per_second": 1985.032, | |
| "eval_steps_per_second": 31.384, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 6.808986663818359, | |
| "learning_rate": 1.263157894736842e-05, | |
| "loss": 0.9812, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "grad_norm": 9.487598419189453, | |
| "learning_rate": 1.2280701754385966e-05, | |
| "loss": 1.0201, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 8.165717124938965, | |
| "learning_rate": 1.192982456140351e-05, | |
| "loss": 0.9722, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "grad_norm": 6.678247928619385, | |
| "learning_rate": 1.1578947368421053e-05, | |
| "loss": 0.9322, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "grad_norm": 7.177236080169678, | |
| "learning_rate": 1.1228070175438597e-05, | |
| "loss": 0.86, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "eval_accuracy": 0.7437417654808959, | |
| "eval_f1_macro": 0.6559797063970845, | |
| "eval_f1_micro": 0.7437417654808959, | |
| "eval_loss": 0.9199798703193665, | |
| "eval_runtime": 0.8169, | |
| "eval_samples_per_second": 1858.172, | |
| "eval_steps_per_second": 29.378, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 6.749398231506348, | |
| "learning_rate": 1.0877192982456142e-05, | |
| "loss": 0.8869, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "grad_norm": 5.880543231964111, | |
| "learning_rate": 1.0526315789473684e-05, | |
| "loss": 0.91, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 7.917410373687744, | |
| "learning_rate": 1.017543859649123e-05, | |
| "loss": 0.9176, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 8.381546974182129, | |
| "learning_rate": 9.824561403508772e-06, | |
| "loss": 0.8586, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "grad_norm": 6.905361652374268, | |
| "learning_rate": 9.473684210526315e-06, | |
| "loss": 0.9646, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_accuracy": 0.7496706192358367, | |
| "eval_f1_macro": 0.6626002653133597, | |
| "eval_f1_micro": 0.7496706192358367, | |
| "eval_loss": 0.8703672885894775, | |
| "eval_runtime": 0.7646, | |
| "eval_samples_per_second": 1985.245, | |
| "eval_steps_per_second": 31.387, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "grad_norm": 7.723905086517334, | |
| "learning_rate": 9.12280701754386e-06, | |
| "loss": 0.9075, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 7.48672342300415, | |
| "learning_rate": 8.771929824561405e-06, | |
| "loss": 0.8439, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "grad_norm": 8.369351387023926, | |
| "learning_rate": 8.421052631578948e-06, | |
| "loss": 0.8985, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 9.050497055053711, | |
| "learning_rate": 8.070175438596492e-06, | |
| "loss": 0.9643, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "grad_norm": 9.8544340133667, | |
| "learning_rate": 7.719298245614036e-06, | |
| "loss": 0.8749, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "eval_accuracy": 0.7707509881422925, | |
| "eval_f1_macro": 0.6960107557890455, | |
| "eval_f1_micro": 0.7707509881422925, | |
| "eval_loss": 0.8367371559143066, | |
| "eval_runtime": 0.8169, | |
| "eval_samples_per_second": 1858.174, | |
| "eval_steps_per_second": 29.378, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 8.317519187927246, | |
| "learning_rate": 7.368421052631579e-06, | |
| "loss": 0.7852, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 8.191059112548828, | |
| "learning_rate": 7.017543859649123e-06, | |
| "loss": 0.8442, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 9.301290512084961, | |
| "learning_rate": 6.666666666666667e-06, | |
| "loss": 0.8622, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 7.775876522064209, | |
| "learning_rate": 6.31578947368421e-06, | |
| "loss": 0.8358, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "grad_norm": 5.711615562438965, | |
| "learning_rate": 5.964912280701755e-06, | |
| "loss": 0.8031, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "eval_accuracy": 0.7766798418972332, | |
| "eval_f1_macro": 0.7068555084926927, | |
| "eval_f1_micro": 0.7766798418972332, | |
| "eval_loss": 0.8124832510948181, | |
| "eval_runtime": 0.8177, | |
| "eval_samples_per_second": 1856.496, | |
| "eval_steps_per_second": 29.352, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.16, | |
| "grad_norm": 6.298779010772705, | |
| "learning_rate": 5.6140350877192985e-06, | |
| "loss": 0.7218, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "grad_norm": 7.6816816329956055, | |
| "learning_rate": 5.263157894736842e-06, | |
| "loss": 0.8172, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 2.26, | |
| "grad_norm": 7.724653244018555, | |
| "learning_rate": 4.912280701754386e-06, | |
| "loss": 0.7422, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 9.366430282592773, | |
| "learning_rate": 4.56140350877193e-06, | |
| "loss": 0.8233, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "grad_norm": 5.840287685394287, | |
| "learning_rate": 4.210526315789474e-06, | |
| "loss": 0.7728, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_accuracy": 0.7786561264822134, | |
| "eval_f1_macro": 0.7085013173330461, | |
| "eval_f1_micro": 0.7786561264822134, | |
| "eval_loss": 0.791221559047699, | |
| "eval_runtime": 0.768, | |
| "eval_samples_per_second": 1976.614, | |
| "eval_steps_per_second": 31.251, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "grad_norm": 6.4323506355285645, | |
| "learning_rate": 3.859649122807018e-06, | |
| "loss": 0.7272, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "grad_norm": 7.345887660980225, | |
| "learning_rate": 3.5087719298245615e-06, | |
| "loss": 0.7775, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 6.521994590759277, | |
| "learning_rate": 3.157894736842105e-06, | |
| "loss": 0.778, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "grad_norm": 9.85383129119873, | |
| "learning_rate": 2.8070175438596493e-06, | |
| "loss": 0.7382, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "grad_norm": 7.1958088874816895, | |
| "learning_rate": 2.456140350877193e-06, | |
| "loss": 0.7164, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "eval_accuracy": 0.7793148880105402, | |
| "eval_f1_macro": 0.706831738723716, | |
| "eval_f1_micro": 0.7793148880105402, | |
| "eval_loss": 0.7828972339630127, | |
| "eval_runtime": 0.8174, | |
| "eval_samples_per_second": 1857.185, | |
| "eval_steps_per_second": 29.363, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.68, | |
| "grad_norm": 8.842488288879395, | |
| "learning_rate": 2.105263157894737e-06, | |
| "loss": 0.7594, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 7.805200099945068, | |
| "learning_rate": 1.7543859649122807e-06, | |
| "loss": 0.7588, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "grad_norm": 5.833488941192627, | |
| "learning_rate": 1.4035087719298246e-06, | |
| "loss": 0.7678, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "grad_norm": 7.358514785766602, | |
| "learning_rate": 1.0526315789473685e-06, | |
| "loss": 0.72, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "grad_norm": 7.7046003341674805, | |
| "learning_rate": 7.017543859649123e-07, | |
| "loss": 0.7533, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "eval_accuracy": 0.782608695652174, | |
| "eval_f1_macro": 0.713683768180215, | |
| "eval_f1_micro": 0.782608695652174, | |
| "eval_loss": 0.777472734451294, | |
| "eval_runtime": 0.7685, | |
| "eval_samples_per_second": 1975.259, | |
| "eval_steps_per_second": 31.229, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 7.136342525482178, | |
| "learning_rate": 3.5087719298245616e-07, | |
| "loss": 0.8063, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 7.667459487915039, | |
| "learning_rate": 0.0, | |
| "loss": 0.7986, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 570, | |
| "total_flos": 1208555069767680.0, | |
| "train_loss": 1.1344423653786642, | |
| "train_runtime": 82.3443, | |
| "train_samples_per_second": 442.435, | |
| "train_steps_per_second": 6.922 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 570, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 50, | |
| "total_flos": 1208555069767680.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |