| { | |
| "best_global_step": 750, | |
| "best_metric": 0.28985536098480225, | |
| "best_model_checkpoint": "siglip2-finetune-full/checkpoint-750", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 750, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.806875, | |
| "eval_loss": 0.4259241223335266, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 116.0639, | |
| "eval_samples_per_second": 68.928, | |
| "eval_steps_per_second": 8.616, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 8.506980895996094, | |
| "learning_rate": 7.142857142857143e-05, | |
| "loss": 0.5089, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.814125, | |
| "eval_loss": 0.41451969742774963, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 115.6232, | |
| "eval_samples_per_second": 69.19, | |
| "eval_steps_per_second": 8.649, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.877375, | |
| "eval_loss": 0.28985536098480225, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 115.8177, | |
| "eval_samples_per_second": 69.074, | |
| "eval_steps_per_second": 8.634, | |
| "step": 750 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 750, | |
| "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": 2.010131899711488e+18, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |