| { | |
| "best_metric": 0.07582072168588638, | |
| "best_model_checkpoint": "clip_base_2_aug/checkpoint-1125", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 1125, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9616666666666667, | |
| "eval_loss": 0.10130487382411957, | |
| "eval_runtime": 209.7054, | |
| "eval_samples_per_second": 114.446, | |
| "eval_steps_per_second": 1.788, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 1.3333333333333333, | |
| "grad_norm": 15.02544116973877, | |
| "learning_rate": 5.813953488372093e-07, | |
| "loss": 0.2009, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9682083333333333, | |
| "eval_loss": 0.08352391421794891, | |
| "eval_runtime": 113.094, | |
| "eval_samples_per_second": 212.213, | |
| "eval_steps_per_second": 3.316, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.6666666666666665, | |
| "grad_norm": 19.05276107788086, | |
| "learning_rate": 1.1627906976744186e-07, | |
| "loss": 0.0774, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9724166666666667, | |
| "eval_loss": 0.07582072168588638, | |
| "eval_runtime": 106.7446, | |
| "eval_samples_per_second": 224.836, | |
| "eval_steps_per_second": 3.513, | |
| "step": 1125 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1125, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "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": 2.273880039240499e+19, | |
| "train_batch_size": 256, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |