| { | |
| "best_metric": 2.693235397338867, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-560m_de_continual-pretrain_100000samples_-1vocab_original_bsz16/checkpoint-10000", | |
| "epoch": 5.495672482483858, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 9e-05, | |
| "loss": 3.0365, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 8e-05, | |
| "loss": 2.6676, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 2.7494328022003174, | |
| "eval_runtime": 85.3446, | |
| "eval_samples_per_second": 9.421, | |
| "eval_steps_per_second": 4.71, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 7e-05, | |
| "loss": 2.4634, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 6e-05, | |
| "loss": 2.3439, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 2.693235397338867, | |
| "eval_runtime": 85.1842, | |
| "eval_samples_per_second": 9.438, | |
| "eval_steps_per_second": 4.719, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 5e-05, | |
| "loss": 2.2031, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "learning_rate": 4e-05, | |
| "loss": 2.0615, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "eval_loss": 2.737623453140259, | |
| "eval_runtime": 85.2178, | |
| "eval_samples_per_second": 9.435, | |
| "eval_steps_per_second": 4.717, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.85, | |
| "learning_rate": 3e-05, | |
| "loss": 1.9567, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "learning_rate": 2e-05, | |
| "loss": 1.7524, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "eval_loss": 2.913102626800537, | |
| "eval_runtime": 85.2676, | |
| "eval_samples_per_second": 9.429, | |
| "eval_steps_per_second": 4.715, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 1e-05, | |
| "loss": 1.6855, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "learning_rate": 0.0, | |
| "loss": 1.4589, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "eval_loss": 3.240140914916992, | |
| "eval_runtime": 85.4709, | |
| "eval_samples_per_second": 9.407, | |
| "eval_steps_per_second": 4.703, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "step": 25000, | |
| "total_flos": 7.43021311206359e+17, | |
| "train_loss": 2.162963125, | |
| "train_runtime": 135519.3536, | |
| "train_samples_per_second": 2.952, | |
| "train_steps_per_second": 0.184 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 6, | |
| "total_flos": 7.43021311206359e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |