| { |
| "best_global_step": 3000, |
| "best_metric": 0.3550371527671814, |
| "best_model_checkpoint": "siglip2-image-classification/checkpoint-3000", |
| "epoch": 2.0, |
| "eval_steps": 500, |
| "global_step": 3000, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.3333333333333333, |
| "grad_norm": 3.479196786880493, |
| "learning_rate": 0.00018490756302521008, |
| "loss": 0.5047473754882813, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.6666666666666666, |
| "grad_norm": 4.285048007965088, |
| "learning_rate": 0.00016810084033613447, |
| "loss": 0.42015097045898436, |
| "step": 1000 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 2.655521869659424, |
| "learning_rate": 0.00015129411764705882, |
| "loss": 0.38898434448242186, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.8312134754211069, |
| "eval_loss": 0.37709423899650574, |
| "eval_model_preparation_time": 0.0019, |
| "eval_runtime": 283.3202, |
| "eval_samples_per_second": 112.943, |
| "eval_steps_per_second": 14.118, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.3333333333333333, |
| "grad_norm": 1.3074887990951538, |
| "learning_rate": 0.0001344873949579832, |
| "loss": 0.3693521728515625, |
| "step": 2000 |
| }, |
| { |
| "epoch": 1.6666666666666665, |
| "grad_norm": 1.2879441976547241, |
| "learning_rate": 0.00011768067226890757, |
| "loss": 0.3484758605957031, |
| "step": 2500 |
| }, |
| { |
| "epoch": 2.0, |
| "grad_norm": 1.5275274515151978, |
| "learning_rate": 0.00010087394957983194, |
| "loss": 0.34061029052734376, |
| "step": 3000 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.8569642801337541, |
| "eval_loss": 0.3550371527671814, |
| "eval_model_preparation_time": 0.0019, |
| "eval_runtime": 282.8579, |
| "eval_samples_per_second": 113.127, |
| "eval_steps_per_second": 14.141, |
| "step": 3000 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 6000, |
| "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": 8.040192576862667e+18, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
