| { | |
| "best_metric": 2.6485073566436768, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-560m_de_continual-pretrain_100000samples_-1vocab_original_bsz8/checkpoint-25000", | |
| "epoch": 2.7477950267893942, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 9e-05, | |
| "loss": 3.2508, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 8e-05, | |
| "loss": 2.8568, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_loss": 2.894188165664673, | |
| "eval_runtime": 86.1381, | |
| "eval_samples_per_second": 9.334, | |
| "eval_steps_per_second": 4.667, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 7e-05, | |
| "loss": 2.7436, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 6e-05, | |
| "loss": 2.5927, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 2.746913433074951, | |
| "eval_runtime": 86.1236, | |
| "eval_samples_per_second": 9.335, | |
| "eval_steps_per_second": 4.668, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 5e-05, | |
| "loss": 2.4376, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 4e-05, | |
| "loss": 2.4191, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_loss": 2.6694421768188477, | |
| "eval_runtime": 86.3384, | |
| "eval_samples_per_second": 9.312, | |
| "eval_steps_per_second": 4.656, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 3e-05, | |
| "loss": 2.3791, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 2e-05, | |
| "loss": 2.168, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 2.6745970249176025, | |
| "eval_runtime": 86.1621, | |
| "eval_samples_per_second": 9.331, | |
| "eval_steps_per_second": 4.666, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 1e-05, | |
| "loss": 2.0728, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 0.0, | |
| "loss": 2.0629, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "eval_loss": 2.6485073566436768, | |
| "eval_runtime": 86.1418, | |
| "eval_samples_per_second": 9.333, | |
| "eval_steps_per_second": 4.667, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "step": 25000, | |
| "total_flos": 3.715050833612636e+17, | |
| "train_loss": 2.498347421875, | |
| "train_runtime": 69506.5729, | |
| "train_samples_per_second": 2.877, | |
| "train_steps_per_second": 0.36 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 3, | |
| "total_flos": 3.715050833612636e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |