| { | |
| "best_metric": 0.3348920941352844, | |
| "best_model_checkpoint": "./vit5_qqp/checkpoint-863", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2589, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5793742757821553, | |
| "grad_norm": 2.493626356124878, | |
| "learning_rate": 4.034376207029741e-05, | |
| "loss": 0.4544, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.3348920941352844, | |
| "eval_runtime": 116.9015, | |
| "eval_samples_per_second": 184.446, | |
| "eval_steps_per_second": 4.619, | |
| "step": 863 | |
| }, | |
| { | |
| "epoch": 1.1587485515643106, | |
| "grad_norm": 0.8472898006439209, | |
| "learning_rate": 3.0687524140594824e-05, | |
| "loss": 0.2056, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.7381228273464657, | |
| "grad_norm": 0.994403600692749, | |
| "learning_rate": 2.103128621089224e-05, | |
| "loss": 0.155, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.3606082797050476, | |
| "eval_runtime": 116.9754, | |
| "eval_samples_per_second": 184.329, | |
| "eval_steps_per_second": 4.616, | |
| "step": 1726 | |
| }, | |
| { | |
| "epoch": 2.317497103128621, | |
| "grad_norm": 0.6376566886901855, | |
| "learning_rate": 1.1375048281189648e-05, | |
| "loss": 0.125, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.8968713789107765, | |
| "grad_norm": 0.6937958598136902, | |
| "learning_rate": 1.7188103514870608e-06, | |
| "loss": 0.1083, | |
| "step": 2500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2589, | |
| "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": 8763161692078080.0, | |
| "train_batch_size": 40, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |