| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.23076923076923, | |
| "eval_steps": 500, | |
| "global_step": 60, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15384615384615385, | |
| "grad_norm": 4.90625, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 2.7975, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.7692307692307693, | |
| "grad_norm": 1.484375, | |
| "learning_rate": 0.0001666666666666667, | |
| "loss": 2.777, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.9230769230769231, | |
| "eval_loss": 2.9954144954681396, | |
| "eval_runtime": 0.6514, | |
| "eval_samples_per_second": 15.352, | |
| "eval_steps_per_second": 1.535, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 3.5625, | |
| "learning_rate": 0.00019730448705798239, | |
| "loss": 2.4096, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.712022304534912, | |
| "eval_runtime": 0.5378, | |
| "eval_samples_per_second": 18.593, | |
| "eval_steps_per_second": 1.859, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 2.3076923076923075, | |
| "grad_norm": 1.25, | |
| "learning_rate": 0.00018660254037844388, | |
| "loss": 2.0954, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 2.9230769230769234, | |
| "eval_loss": 2.6308600902557373, | |
| "eval_runtime": 0.679, | |
| "eval_samples_per_second": 14.728, | |
| "eval_steps_per_second": 1.473, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 3.076923076923077, | |
| "grad_norm": 1.09375, | |
| "learning_rate": 0.0001686241637868734, | |
| "loss": 1.8931, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.8461538461538463, | |
| "grad_norm": 1.8828125, | |
| "learning_rate": 0.00014487991802004623, | |
| "loss": 1.7235, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 2.5914011001586914, | |
| "eval_runtime": 0.541, | |
| "eval_samples_per_second": 18.486, | |
| "eval_steps_per_second": 1.849, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.615384615384615, | |
| "grad_norm": 0.421875, | |
| "learning_rate": 0.00011736481776669306, | |
| "loss": 1.5996, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 4.923076923076923, | |
| "eval_loss": 2.5729358196258545, | |
| "eval_runtime": 0.6344, | |
| "eval_samples_per_second": 15.764, | |
| "eval_steps_per_second": 1.576, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 5.384615384615385, | |
| "grad_norm": 0.90625, | |
| "learning_rate": 8.839070858747697e-05, | |
| "loss": 1.5071, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 2.5703837871551514, | |
| "eval_runtime": 0.5366, | |
| "eval_samples_per_second": 18.636, | |
| "eval_steps_per_second": 1.864, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 6.153846153846154, | |
| "grad_norm": 0.310546875, | |
| "learning_rate": 6.039202339608432e-05, | |
| "loss": 1.4579, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 6.923076923076923, | |
| "grad_norm": 0.498046875, | |
| "learning_rate": 3.5721239031346066e-05, | |
| "loss": 1.4254, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 6.923076923076923, | |
| "eval_loss": 2.570425510406494, | |
| "eval_runtime": 0.6976, | |
| "eval_samples_per_second": 14.334, | |
| "eval_steps_per_second": 1.433, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 7.6923076923076925, | |
| "grad_norm": 0.353515625, | |
| "learning_rate": 1.6451218858706374e-05, | |
| "loss": 1.3939, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 2.5644469261169434, | |
| "eval_runtime": 0.5372, | |
| "eval_samples_per_second": 18.616, | |
| "eval_steps_per_second": 1.862, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 8.461538461538462, | |
| "grad_norm": 0.84765625, | |
| "learning_rate": 4.20104876845111e-06, | |
| "loss": 1.3833, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 8.923076923076923, | |
| "eval_loss": 2.5641579627990723, | |
| "eval_runtime": 0.6811, | |
| "eval_samples_per_second": 14.681, | |
| "eval_steps_per_second": 1.468, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 9.23076923076923, | |
| "grad_norm": 0.357421875, | |
| "learning_rate": 0.0, | |
| "loss": 1.3864, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 9.23076923076923, | |
| "eval_loss": 2.564016819000244, | |
| "eval_runtime": 0.5352, | |
| "eval_samples_per_second": 18.685, | |
| "eval_steps_per_second": 1.868, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 9.23076923076923, | |
| "step": 60, | |
| "total_flos": 3.5242400385335296e+16, | |
| "train_loss": 1.7546881834665935, | |
| "train_runtime": 212.0463, | |
| "train_samples_per_second": 14.478, | |
| "train_steps_per_second": 0.283 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 60, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 3.5242400385335296e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |