| { | |
| "best_metric": 0.7333333333333333, | |
| "best_model_checkpoint": "distilbert-base-multilingual-cased-hyper-matt/run-esi21ugw/checkpoint-50", | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 50, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.9570593237876892, | |
| "learning_rate": 2.1284157243016352e-05, | |
| "loss": 0.6082, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 3.792499542236328, | |
| "learning_rate": 1.5963117932262263e-05, | |
| "loss": 0.5307, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 3.5518383979797363, | |
| "learning_rate": 1.0642078621508176e-05, | |
| "loss": 0.4009, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 5.375403881072998, | |
| "learning_rate": 5.321039310754088e-06, | |
| "loss": 0.3688, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.561539649963379, | |
| "learning_rate": 0.0, | |
| "loss": 0.3779, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.84, | |
| "eval_f1": 0.7333333333333333, | |
| "eval_loss": 0.36742866039276123, | |
| "eval_precision": 0.7521367521367521, | |
| "eval_recall": 0.7154471544715447, | |
| "eval_runtime": 1.5461, | |
| "eval_samples_per_second": 258.717, | |
| "eval_steps_per_second": 16.17, | |
| "step": 50 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 50, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 211815370450944.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": { | |
| "_wandb": {}, | |
| "assignments": {}, | |
| "learning_rate": 2.660519655377044e-05, | |
| "metric": "eval/loss", | |
| "num_train_epochs": 1, | |
| "per_device_train_batch_size": 32, | |
| "seed": 9 | |
| } | |
| } | |