| { | |
| "best_global_step": 50, | |
| "best_metric": 0.8024441003799438, | |
| "best_model_checkpoint": "/home/ricoiban/GEMMA/mnlp_chatsplaining/safety_model_output/checkpoint-50", | |
| "epoch": 1.0207253886010363, | |
| "eval_steps": 50, | |
| "global_step": 50, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.20725388601036268, | |
| "grad_norm": 4.4248785972595215, | |
| "learning_rate": 0.0005899999999999999, | |
| "loss": 8.2557, | |
| "mean_token_accuracy": 0.384375, | |
| "num_tokens": 21984.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.41450777202072536, | |
| "grad_norm": 1.9055719375610352, | |
| "learning_rate": 0.000565, | |
| "loss": 0.7267, | |
| "mean_token_accuracy": 0.634375, | |
| "num_tokens": 44111.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.6217616580310881, | |
| "grad_norm": 2.149226427078247, | |
| "learning_rate": 0.00054, | |
| "loss": 0.6712, | |
| "mean_token_accuracy": 0.671875, | |
| "num_tokens": 68120.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.8290155440414507, | |
| "grad_norm": 1.9442976713180542, | |
| "learning_rate": 0.0005149999999999999, | |
| "loss": 0.7079, | |
| "mean_token_accuracy": 0.678125, | |
| "num_tokens": 89669.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.0207253886010363, | |
| "grad_norm": 67.3375473022461, | |
| "learning_rate": 0.00049, | |
| "loss": 0.7584, | |
| "mean_token_accuracy": 0.6182432432432432, | |
| "num_tokens": 111297.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.0207253886010363, | |
| "eval_loss": 0.8024441003799438, | |
| "eval_mean_token_accuracy": 0.5, | |
| "eval_num_tokens": 111297.0, | |
| "eval_runtime": 0.1286, | |
| "eval_samples_per_second": 7.775, | |
| "eval_steps_per_second": 7.775, | |
| "step": 50 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 245, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 424335605760000.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |