| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.15384615384615385, | |
| "eval_steps": 10, | |
| "global_step": 50, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.006153846153846154, | |
| "eval_loss": 1.845949411392212, | |
| "eval_runtime": 1.9761, | |
| "eval_samples_per_second": 57.69, | |
| "eval_steps_per_second": 3.036, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.03076923076923077, | |
| "grad_norm": 29.46457290649414, | |
| "learning_rate": 7.692307692307694e-07, | |
| "loss": 1.944, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.03076923076923077, | |
| "eval_loss": 1.7965577840805054, | |
| "eval_runtime": 1.9506, | |
| "eval_samples_per_second": 58.444, | |
| "eval_steps_per_second": 3.076, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06153846153846154, | |
| "grad_norm": 14.49207878112793, | |
| "learning_rate": 1.5384615384615387e-06, | |
| "loss": 1.6999, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06153846153846154, | |
| "eval_loss": 1.5032916069030762, | |
| "eval_runtime": 1.9599, | |
| "eval_samples_per_second": 58.168, | |
| "eval_steps_per_second": 3.061, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09230769230769231, | |
| "grad_norm": 11.355466842651367, | |
| "learning_rate": 2.307692307692308e-06, | |
| "loss": 1.1499, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.09230769230769231, | |
| "eval_loss": 0.8391174674034119, | |
| "eval_runtime": 1.9437, | |
| "eval_samples_per_second": 58.65, | |
| "eval_steps_per_second": 3.087, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12307692307692308, | |
| "grad_norm": 5.648468017578125, | |
| "learning_rate": 3.0769230769230774e-06, | |
| "loss": 0.7194, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.12307692307692308, | |
| "eval_loss": 0.6716201305389404, | |
| "eval_runtime": 1.9517, | |
| "eval_samples_per_second": 58.411, | |
| "eval_steps_per_second": 3.074, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.15384615384615385, | |
| "grad_norm": 5.752841949462891, | |
| "learning_rate": 3.846153846153847e-06, | |
| "loss": 0.6811, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.15384615384615385, | |
| "eval_loss": 0.6289324164390564, | |
| "eval_runtime": 1.952, | |
| "eval_samples_per_second": 58.401, | |
| "eval_steps_per_second": 3.074, | |
| "step": 50 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 2600, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 50, | |
| "total_flos": 4268289777729536.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |