| { | |
| "best_metric": 1.7216796875, | |
| "best_model_checkpoint": "/home/zhengxinyong/outputs/bloom-1b7_ru_continual-pretrain_100000samples_-1vocab_original/checkpoint-25000", | |
| "epoch": 1.4637002341920375, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 0.0001, | |
| "loss": 2.2537, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9859, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "eval_loss": 1.9208984375, | |
| "eval_runtime": 7.2654, | |
| "eval_samples_per_second": 187.601, | |
| "eval_steps_per_second": 23.536, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 0.0001, | |
| "loss": 1.896, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 0.0001, | |
| "loss": 1.853, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_loss": 1.82421875, | |
| "eval_runtime": 7.2655, | |
| "eval_samples_per_second": 187.599, | |
| "eval_steps_per_second": 23.536, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8152, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 0.0001, | |
| "loss": 1.782, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_loss": 1.7744140625, | |
| "eval_runtime": 7.2563, | |
| "eval_samples_per_second": 187.838, | |
| "eval_steps_per_second": 23.566, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7434, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6356, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_loss": 1.744140625, | |
| "eval_runtime": 7.2585, | |
| "eval_samples_per_second": 187.781, | |
| "eval_steps_per_second": 23.559, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6475, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6398, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "eval_loss": 1.7216796875, | |
| "eval_runtime": 7.2617, | |
| "eval_samples_per_second": 187.696, | |
| "eval_steps_per_second": 23.548, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "step": 25000, | |
| "total_flos": 1.4851359536421274e+18, | |
| "train_loss": 1.8252051953125, | |
| "train_runtime": 6127.1413, | |
| "train_samples_per_second": 32.642, | |
| "train_steps_per_second": 4.08 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 2, | |
| "total_flos": 1.4851359536421274e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |