| { |
| "best_global_step": 2756, |
| "best_metric": 1.0952644348144531, |
| "best_model_checkpoint": "siglip2-finetune-full/checkpoint-2756", |
| "epoch": 2.0, |
| "eval_steps": 500, |
| "global_step": 2756, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.36284470246734396, |
| "grad_norm": 5.684016704559326, |
| "learning_rate": 0.00017796278158667975, |
| "loss": 1.841, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.7256894049346879, |
| "grad_norm": 4.77960205078125, |
| "learning_rate": 0.0001534769833496572, |
| "loss": 1.4093, |
| "step": 1000 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.6121346886912326, |
| "eval_loss": 1.3363316059112549, |
| "eval_model_preparation_time": 0.0023, |
| "eval_runtime": 540.1365, |
| "eval_samples_per_second": 81.594, |
| "eval_steps_per_second": 10.199, |
| "step": 1378 |
| }, |
| { |
| "epoch": 1.0885341074020318, |
| "grad_norm": 4.179084300994873, |
| "learning_rate": 0.00012899118511263468, |
| "loss": 1.3033, |
| "step": 1500 |
| }, |
| { |
| "epoch": 1.4513788098693758, |
| "grad_norm": 4.6432366371154785, |
| "learning_rate": 0.00010450538687561216, |
| "loss": 1.2174, |
| "step": 2000 |
| }, |
| { |
| "epoch": 1.8142235123367199, |
| "grad_norm": 2.942708969116211, |
| "learning_rate": 8.001958863858962e-05, |
| "loss": 1.1675, |
| "step": 2500 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.6670902160101652, |
| "eval_loss": 1.0952644348144531, |
| "eval_model_preparation_time": 0.0023, |
| "eval_runtime": 535.0051, |
| "eval_samples_per_second": 82.377, |
| "eval_steps_per_second": 10.297, |
| "step": 2756 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 4134, |
| "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": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 7.385238068390461e+18, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
