| { | |
| "best_global_step": 2484, | |
| "best_metric": 0.43748077750205994, | |
| "best_model_checkpoint": "siglip2-finetune-full/checkpoint-2484", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2484, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6038647342995169, | |
| "grad_norm": 3.9172444343566895, | |
| "learning_rate": 0.00016302382908792113, | |
| "loss": 1.1039, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7029792697201979, | |
| "eval_loss": 0.7443631291389465, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 481.4814, | |
| "eval_samples_per_second": 55.003, | |
| "eval_steps_per_second": 6.877, | |
| "step": 828 | |
| }, | |
| { | |
| "epoch": 1.2077294685990339, | |
| "grad_norm": 4.54465913772583, | |
| "learning_rate": 0.00012193919474116682, | |
| "loss": 0.8238, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.8115942028985508, | |
| "grad_norm": 9.120803833007812, | |
| "learning_rate": 8.08545603944125e-05, | |
| "loss": 0.71, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.785258467696258, | |
| "eval_loss": 0.5553735494613647, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 482.7837, | |
| "eval_samples_per_second": 54.855, | |
| "eval_steps_per_second": 6.858, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 2.4154589371980677, | |
| "grad_norm": 5.409132957458496, | |
| "learning_rate": 3.976992604765818e-05, | |
| "loss": 0.5791, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8295510327379828, | |
| "eval_loss": 0.43748077750205994, | |
| "eval_model_preparation_time": 0.0023, | |
| "eval_runtime": 481.8468, | |
| "eval_samples_per_second": 54.961, | |
| "eval_steps_per_second": 6.871, | |
| "step": 2484 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2484, | |
| "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": 6.654455928125172e+18, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |