| { | |
| "best_metric": 0.7644787644787645, | |
| "best_model_checkpoint": "distilbert-base-multilingual-cased-hyper-matt/run-xxn3kofd/checkpoint-400", | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 400, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 2.57446026802063, | |
| "learning_rate": 8.430421587841344e-05, | |
| "loss": 0.6104, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 72.49688720703125, | |
| "learning_rate": 8.214256931742847e-05, | |
| "loss": 0.4612, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 3.086334705352783, | |
| "learning_rate": 7.998092275644352e-05, | |
| "loss": 0.4366, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 1.2897779941558838, | |
| "learning_rate": 7.781927619545856e-05, | |
| "loss": 0.4857, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 2.7454018592834473, | |
| "learning_rate": 7.56576296344736e-05, | |
| "loss": 0.4842, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 13.36482048034668, | |
| "learning_rate": 7.349598307348864e-05, | |
| "loss": 0.6072, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 1.689212441444397, | |
| "learning_rate": 7.133433651250367e-05, | |
| "loss": 0.3871, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1.120545744895935, | |
| "learning_rate": 6.917268995151872e-05, | |
| "loss": 0.5613, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 12.049583435058594, | |
| "learning_rate": 6.701104339053376e-05, | |
| "loss": 0.5383, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 1.9796591997146606, | |
| "learning_rate": 6.48493968295488e-05, | |
| "loss": 0.5167, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 1.916353464126587, | |
| "learning_rate": 6.268775026856383e-05, | |
| "loss": 0.5366, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 3.923786163330078, | |
| "learning_rate": 6.052610370757888e-05, | |
| "loss": 0.4549, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 7.377273082733154, | |
| "learning_rate": 5.836445714659392e-05, | |
| "loss": 0.3999, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 8.601259231567383, | |
| "learning_rate": 5.6202810585608965e-05, | |
| "loss": 0.2955, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 21.193166732788086, | |
| "learning_rate": 5.4041164024624e-05, | |
| "loss": 0.4126, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 13.54879379272461, | |
| "learning_rate": 5.187951746363904e-05, | |
| "loss": 0.3501, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 2.4327757358551025, | |
| "learning_rate": 4.9717870902654076e-05, | |
| "loss": 0.4196, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 4.049587249755859, | |
| "learning_rate": 4.7556224341669126e-05, | |
| "loss": 0.2942, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 14.283734321594238, | |
| "learning_rate": 4.539457778068416e-05, | |
| "loss": 0.4683, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.0148279666900635, | |
| "learning_rate": 4.32329312196992e-05, | |
| "loss": 0.4988, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.705, | |
| "eval_f1": 0.6685393258426966, | |
| "eval_loss": 0.5600033402442932, | |
| "eval_precision": 0.5107296137339056, | |
| "eval_recall": 0.967479674796748, | |
| "eval_runtime": 1.5406, | |
| "eval_samples_per_second": 259.637, | |
| "eval_steps_per_second": 16.227, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 13.742440223693848, | |
| "learning_rate": 4.107128465871424e-05, | |
| "loss": 0.454, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 5.356462001800537, | |
| "learning_rate": 3.890963809772928e-05, | |
| "loss": 0.3232, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 8.677199363708496, | |
| "learning_rate": 3.674799153674432e-05, | |
| "loss": 0.3169, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 4.404296875, | |
| "learning_rate": 3.458634497575936e-05, | |
| "loss": 0.264, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 2.6422922611236572, | |
| "learning_rate": 3.24246984147744e-05, | |
| "loss": 0.1895, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 1.4365283250808716, | |
| "learning_rate": 3.026305185378944e-05, | |
| "loss": 0.3452, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "grad_norm": 5.857382297515869, | |
| "learning_rate": 2.8101405292804482e-05, | |
| "loss": 0.2193, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 4.946323394775391, | |
| "learning_rate": 2.593975873181952e-05, | |
| "loss": 0.6003, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 3.101881504058838, | |
| "learning_rate": 2.3778112170834563e-05, | |
| "loss": 0.3291, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 11.071148872375488, | |
| "learning_rate": 2.16164656098496e-05, | |
| "loss": 0.3606, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 3.46047306060791, | |
| "learning_rate": 1.945481904886464e-05, | |
| "loss": 0.2745, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 7.391355991363525, | |
| "learning_rate": 1.729317248787968e-05, | |
| "loss": 0.5095, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 0.8834757208824158, | |
| "learning_rate": 1.513152592689472e-05, | |
| "loss": 0.247, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "grad_norm": 0.27169790863990784, | |
| "learning_rate": 1.296987936590976e-05, | |
| "loss": 0.4244, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 1.2865382432937622, | |
| "learning_rate": 1.08082328049248e-05, | |
| "loss": 0.4902, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 3.9220056533813477, | |
| "learning_rate": 8.64658624393984e-06, | |
| "loss": 0.3548, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.0294697284698486, | |
| "learning_rate": 6.48493968295488e-06, | |
| "loss": 0.3243, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 9.069091796875, | |
| "learning_rate": 4.32329312196992e-06, | |
| "loss": 0.3318, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 3.9282755851745605, | |
| "learning_rate": 2.16164656098496e-06, | |
| "loss": 0.221, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.1504770517349243, | |
| "learning_rate": 0.0, | |
| "loss": 0.365, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8475, | |
| "eval_f1": 0.7644787644787645, | |
| "eval_loss": 0.35866764187812805, | |
| "eval_precision": 0.7279411764705882, | |
| "eval_recall": 0.8048780487804879, | |
| "eval_runtime": 1.5614, | |
| "eval_samples_per_second": 256.176, | |
| "eval_steps_per_second": 16.011, | |
| "step": 400 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 400, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 423630740901888.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": { | |
| "_wandb": {}, | |
| "assignments": {}, | |
| "learning_rate": 8.64658624393984e-05, | |
| "metric": "eval/loss", | |
| "num_train_epochs": 2, | |
| "per_device_train_batch_size": 8, | |
| "seed": 7 | |
| } | |
| } | |