| { | |
| "best_metric": 3.296297788619995, | |
| "best_model_checkpoint": "output/architects/checkpoint-54", | |
| "epoch": 3.0, | |
| "global_step": 54, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 0.00011269523002449659, | |
| "loss": 3.5266, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 5.668773501204858e-05, | |
| "loss": 3.2821, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 9.190657300387505e-06, | |
| "loss": 3.0804, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 3.374046564102173, | |
| "eval_runtime": 0.3613, | |
| "eval_samples_per_second": 74.727, | |
| "eval_steps_per_second": 11.071, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 4.137086214086682e-06, | |
| "loss": 3.1672, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 4.513741816785908e-05, | |
| "loss": 3.1396, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 0.00010290000000000001, | |
| "loss": 2.9984, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "learning_rate": 0.00013615781185663748, | |
| "loss": 2.8167, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 3.3096883296966553, | |
| "eval_runtime": 0.3607, | |
| "eval_samples_per_second": 74.864, | |
| "eval_steps_per_second": 11.091, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 0.0001211506487979619, | |
| "loss": 2.813, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 6.860000000000001e-05, | |
| "loss": 2.6963, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 1.6049351202038163e-05, | |
| "loss": 2.5468, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 3.296297788619995, | |
| "eval_runtime": 0.3624, | |
| "eval_samples_per_second": 74.5, | |
| "eval_steps_per_second": 11.037, | |
| "step": 54 | |
| } | |
| ], | |
| "max_steps": 900, | |
| "num_train_epochs": 50, | |
| "total_flos": 53695512576000.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |