| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.011534025374855825, | |
| "eval_steps": 3, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0011534025374855825, | |
| "grad_norm": 7.599985122680664, | |
| "learning_rate": 2e-05, | |
| "loss": 5.2804, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0011534025374855825, | |
| "eval_loss": 1.365613341331482, | |
| "eval_runtime": 11.0665, | |
| "eval_samples_per_second": 32.982, | |
| "eval_steps_per_second": 16.536, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.002306805074971165, | |
| "grad_norm": 7.2444233894348145, | |
| "learning_rate": 4e-05, | |
| "loss": 5.2227, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0034602076124567475, | |
| "grad_norm": 7.416990756988525, | |
| "learning_rate": 6e-05, | |
| "loss": 5.297, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.0034602076124567475, | |
| "eval_loss": 1.3492553234100342, | |
| "eval_runtime": 10.9717, | |
| "eval_samples_per_second": 33.267, | |
| "eval_steps_per_second": 16.679, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00461361014994233, | |
| "grad_norm": 7.789440631866455, | |
| "learning_rate": 8e-05, | |
| "loss": 5.896, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.0057670126874279125, | |
| "grad_norm": 11.984971046447754, | |
| "learning_rate": 0.0001, | |
| "loss": 5.8886, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.006920415224913495, | |
| "grad_norm": 5.711300373077393, | |
| "learning_rate": 0.00012, | |
| "loss": 4.5868, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.006920415224913495, | |
| "eval_loss": 1.2016804218292236, | |
| "eval_runtime": 11.0481, | |
| "eval_samples_per_second": 33.037, | |
| "eval_steps_per_second": 16.564, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.008073817762399077, | |
| "grad_norm": 7.23464822769165, | |
| "learning_rate": 0.00014, | |
| "loss": 4.4143, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.00922722029988466, | |
| "grad_norm": 6.925653457641602, | |
| "learning_rate": 0.00016, | |
| "loss": 4.2078, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.010380622837370242, | |
| "grad_norm": 6.105075836181641, | |
| "learning_rate": 0.00018, | |
| "loss": 4.3888, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.010380622837370242, | |
| "eval_loss": 0.9751999974250793, | |
| "eval_runtime": 11.0821, | |
| "eval_samples_per_second": 32.936, | |
| "eval_steps_per_second": 16.513, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.011534025374855825, | |
| "grad_norm": 5.90463924407959, | |
| "learning_rate": 0.0002, | |
| "loss": 4.4459, | |
| "step": 10 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 10, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 3, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 917953377730560.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |