| { | |
| "best_global_step": 1785, | |
| "best_metric": 0.2768375873565674, | |
| "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1785", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 1785, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.8403361344537815, | |
| "grad_norm": 37.12611770629883, | |
| "learning_rate": 4.034334763948498e-06, | |
| "loss": 0.757, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8265671013883046, | |
| "eval_loss": 0.44112637639045715, | |
| "eval_model_preparation_time": 0.0026, | |
| "eval_runtime": 308.0267, | |
| "eval_samples_per_second": 61.735, | |
| "eval_steps_per_second": 7.717, | |
| "step": 595 | |
| }, | |
| { | |
| "epoch": 1.680672268907563, | |
| "grad_norm": 40.48983383178711, | |
| "learning_rate": 2.96137339055794e-06, | |
| "loss": 0.4737, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8392406394615061, | |
| "eval_loss": 0.41642823815345764, | |
| "eval_model_preparation_time": 0.0026, | |
| "eval_runtime": 305.306, | |
| "eval_samples_per_second": 62.285, | |
| "eval_steps_per_second": 7.786, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 2.5210084033613445, | |
| "grad_norm": 30.135860443115234, | |
| "learning_rate": 1.8884120171673823e-06, | |
| "loss": 0.3811, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.889724442574674, | |
| "eval_loss": 0.2768375873565674, | |
| "eval_model_preparation_time": 0.0026, | |
| "eval_runtime": 304.3734, | |
| "eval_samples_per_second": 62.476, | |
| "eval_steps_per_second": 7.809, | |
| "step": 1785 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2380, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.778202391316251e+18, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |