| { | |
| "best_metric": 2.67578125, | |
| "best_model_checkpoint": "/home/zhengxinyong/outputs/bloom-1b7_de_continual-pretrain_100000samples_-1vocab_original/checkpoint-25000", | |
| "epoch": 2.7475546763380594, | |
| "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": 0.0001, | |
| "loss": 3.1493, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 0.0001, | |
| "loss": 2.8195, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_loss": 2.85546875, | |
| "eval_runtime": 4.3012, | |
| "eval_samples_per_second": 186.927, | |
| "eval_steps_per_second": 23.482, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 0.0001, | |
| "loss": 2.7071, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 0.0001, | |
| "loss": 2.5574, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 2.744140625, | |
| "eval_runtime": 4.3032, | |
| "eval_samples_per_second": 186.837, | |
| "eval_steps_per_second": 23.471, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 0.0001, | |
| "loss": 2.4098, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0001, | |
| "loss": 2.4193, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_loss": 2.685546875, | |
| "eval_runtime": 4.2993, | |
| "eval_samples_per_second": 187.007, | |
| "eval_steps_per_second": 23.492, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 0.0001, | |
| "loss": 2.413, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 0.0001, | |
| "loss": 2.173, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 2.716796875, | |
| "eval_runtime": 4.302, | |
| "eval_samples_per_second": 186.892, | |
| "eval_steps_per_second": 23.478, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 0.0001, | |
| "loss": 2.1243, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 0.0001, | |
| "loss": 2.1665, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "eval_loss": 2.67578125, | |
| "eval_runtime": 4.3027, | |
| "eval_samples_per_second": 186.858, | |
| "eval_steps_per_second": 23.473, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "step": 25000, | |
| "total_flos": 1.4851359536421274e+18, | |
| "train_loss": 2.49391578125, | |
| "train_runtime": 6130.9938, | |
| "train_samples_per_second": 32.621, | |
| "train_steps_per_second": 4.078 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 3, | |
| "total_flos": 1.4851359536421274e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |