| { | |
| "best_metric": 2.763396739959717, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-560m_de_continual-pretrain_100000samples_-1vocab_original_bsz2/checkpoint-25000", | |
| "epoch": 0.6869075422448139, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 9e-05, | |
| "loss": 3.7384, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 8e-05, | |
| "loss": 3.3183, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "eval_loss": 3.2969818115234375, | |
| "eval_runtime": 84.13, | |
| "eval_samples_per_second": 9.557, | |
| "eval_steps_per_second": 4.778, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 7e-05, | |
| "loss": 3.1486, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 6e-05, | |
| "loss": 3.0317, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "eval_loss": 3.0825796127319336, | |
| "eval_runtime": 84.1046, | |
| "eval_samples_per_second": 9.56, | |
| "eval_steps_per_second": 4.78, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 5e-05, | |
| "loss": 2.944, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 4e-05, | |
| "loss": 2.8657, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "eval_loss": 2.939875841140747, | |
| "eval_runtime": 84.1494, | |
| "eval_samples_per_second": 9.554, | |
| "eval_steps_per_second": 4.777, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3e-05, | |
| "loss": 2.8046, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2e-05, | |
| "loss": 2.7595, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_loss": 2.832326650619507, | |
| "eval_runtime": 84.1324, | |
| "eval_samples_per_second": 9.556, | |
| "eval_steps_per_second": 4.778, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1e-05, | |
| "loss": 2.7111, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 0.0, | |
| "loss": 2.6609, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "eval_loss": 2.763396739959717, | |
| "eval_runtime": 84.0579, | |
| "eval_samples_per_second": 9.565, | |
| "eval_steps_per_second": 4.782, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "step": 25000, | |
| "total_flos": 9.28706985984e+16, | |
| "train_loss": 2.99828, | |
| "train_runtime": 18720.6702, | |
| "train_samples_per_second": 2.671, | |
| "train_steps_per_second": 1.335 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 1, | |
| "total_flos": 9.28706985984e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |