| { | |
| "best_metric": 44.04121981988171, | |
| "best_model_checkpoint": "/root/turkic_qa/ru_uzn_models/orig_uzn_roberta_base_model/checkpoint-5430", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 5430, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 543, | |
| "train_exact_match": 6.993006993006993, | |
| "train_f1": 14.72956145700345, | |
| "train_runtime": 8.9919, | |
| "train_samples_per_second": 114.102, | |
| "train_steps_per_second": 4.115 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 22.987104415893555, | |
| "learning_rate": 5e-06, | |
| "loss": 4.7716, | |
| "step": 543 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 6.65625, | |
| "eval_f1": 13.460548189578436, | |
| "eval_runtime": 28.875, | |
| "eval_samples_per_second": 112.935, | |
| "eval_steps_per_second": 4.052, | |
| "step": 543 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1086, | |
| "train_exact_match": 12.987012987012987, | |
| "train_f1": 21.035135926338754, | |
| "train_runtime": 9.0115, | |
| "train_samples_per_second": 112.523, | |
| "train_steps_per_second": 4.106 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 27.96239471435547, | |
| "learning_rate": 1e-05, | |
| "loss": 3.7695, | |
| "step": 1086 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 10.71875, | |
| "eval_f1": 17.87426936207211, | |
| "eval_runtime": 28.7945, | |
| "eval_samples_per_second": 113.251, | |
| "eval_steps_per_second": 4.063, | |
| "step": 1086 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1629, | |
| "train_exact_match": 25.174825174825173, | |
| "train_f1": 34.8258899586999, | |
| "train_runtime": 9.0183, | |
| "train_samples_per_second": 113.214, | |
| "train_steps_per_second": 4.103 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 38.839942932128906, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 3.4391, | |
| "step": 1629 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 15.84375, | |
| "eval_f1": 25.41741224618572, | |
| "eval_runtime": 29.0024, | |
| "eval_samples_per_second": 112.439, | |
| "eval_steps_per_second": 4.034, | |
| "step": 1629 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2172, | |
| "train_exact_match": 29.97002997002997, | |
| "train_f1": 41.494806937045226, | |
| "train_runtime": 9.0706, | |
| "train_samples_per_second": 112.671, | |
| "train_steps_per_second": 4.079 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 35.55900955200195, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 2.9975, | |
| "step": 2172 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 20.4375, | |
| "eval_f1": 31.800313776639523, | |
| "eval_runtime": 28.9344, | |
| "eval_samples_per_second": 112.703, | |
| "eval_steps_per_second": 4.044, | |
| "step": 2172 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 2715, | |
| "train_exact_match": 38.86113886113886, | |
| "train_f1": 51.45493720149751, | |
| "train_runtime": 9.0591, | |
| "train_samples_per_second": 112.483, | |
| "train_steps_per_second": 4.084 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 60.02823257446289, | |
| "learning_rate": 6.25e-06, | |
| "loss": 2.6357, | |
| "step": 2715 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 23.65625, | |
| "eval_f1": 36.81284091086545, | |
| "eval_runtime": 28.8941, | |
| "eval_samples_per_second": 112.861, | |
| "eval_steps_per_second": 4.049, | |
| "step": 2715 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 3258, | |
| "train_exact_match": 47.752247752247754, | |
| "train_f1": 60.68401873806664, | |
| "train_runtime": 8.9792, | |
| "train_samples_per_second": 113.373, | |
| "train_steps_per_second": 4.121 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 37.850128173828125, | |
| "learning_rate": 5e-06, | |
| "loss": 2.3155, | |
| "step": 3258 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 26.125, | |
| "eval_f1": 40.40349142174044, | |
| "eval_runtime": 29.0794, | |
| "eval_samples_per_second": 112.141, | |
| "eval_steps_per_second": 4.023, | |
| "step": 3258 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 3801, | |
| "train_exact_match": 50.44955044955045, | |
| "train_f1": 64.76949286642217, | |
| "train_runtime": 8.8774, | |
| "train_samples_per_second": 113.547, | |
| "train_steps_per_second": 4.055 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 51.20785140991211, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 2.0695, | |
| "step": 3801 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 27.8125, | |
| "eval_f1": 42.61264490274304, | |
| "eval_runtime": 29.0203, | |
| "eval_samples_per_second": 112.37, | |
| "eval_steps_per_second": 4.032, | |
| "step": 3801 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 4344, | |
| "train_exact_match": 51.34865134865135, | |
| "train_f1": 65.11930431582645, | |
| "train_runtime": 8.9923, | |
| "train_samples_per_second": 113.208, | |
| "train_steps_per_second": 4.115 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 56.128509521484375, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.8975, | |
| "step": 4344 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 29.125, | |
| "eval_f1": 43.50290837664036, | |
| "eval_runtime": 28.9786, | |
| "eval_samples_per_second": 112.531, | |
| "eval_steps_per_second": 4.037, | |
| "step": 4344 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 4887, | |
| "train_exact_match": 57.24275724275724, | |
| "train_f1": 69.40786106942967, | |
| "train_runtime": 9.0219, | |
| "train_samples_per_second": 112.615, | |
| "train_steps_per_second": 4.101 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 60.7649040222168, | |
| "learning_rate": 1.25e-06, | |
| "loss": 1.7833, | |
| "step": 4887 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 29.1875, | |
| "eval_f1": 43.93033202135494, | |
| "eval_runtime": 29.1076, | |
| "eval_samples_per_second": 112.033, | |
| "eval_steps_per_second": 4.02, | |
| "step": 4887 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5430, | |
| "train_exact_match": 58.04195804195804, | |
| "train_f1": 70.9609187667352, | |
| "train_runtime": 9.0761, | |
| "train_samples_per_second": 112.383, | |
| "train_steps_per_second": 4.077 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 42.582576751708984, | |
| "learning_rate": 0.0, | |
| "loss": 1.7125, | |
| "step": 5430 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 29.1875, | |
| "eval_f1": 44.04121981988171, | |
| "eval_runtime": 28.9802, | |
| "eval_samples_per_second": 112.525, | |
| "eval_steps_per_second": 4.037, | |
| "step": 5430 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5430, | |
| "total_flos": 1.487387554039296e+16, | |
| "train_loss": 2.7391692272207355, | |
| "train_runtime": 1969.257, | |
| "train_samples_per_second": 77.08, | |
| "train_steps_per_second": 2.757 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5430, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.487387554039296e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |