| { | |
| "best_metric": 1.734375, | |
| "best_model_checkpoint": "/home/zhengxinyong/outputs/bloom-1b1_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": 2.2756, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 0.0001, | |
| "loss": 2.0005, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "eval_loss": 1.9326171875, | |
| "eval_runtime": 5.8642, | |
| "eval_samples_per_second": 232.428, | |
| "eval_steps_per_second": 29.16, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 0.0001, | |
| "loss": 1.9088, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8653, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_loss": 1.8349609375, | |
| "eval_runtime": 5.8764, | |
| "eval_samples_per_second": 231.946, | |
| "eval_steps_per_second": 29.1, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 0.0001, | |
| "loss": 1.8281, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7953, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_loss": 1.7861328125, | |
| "eval_runtime": 5.856, | |
| "eval_samples_per_second": 232.753, | |
| "eval_steps_per_second": 29.201, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "learning_rate": 0.0001, | |
| "loss": 1.7593, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6602, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "eval_loss": 1.7548828125, | |
| "eval_runtime": 5.8587, | |
| "eval_samples_per_second": 232.645, | |
| "eval_steps_per_second": 29.187, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 0.0001, | |
| "loss": 1.672, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 0.0001, | |
| "loss": 1.6639, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "eval_loss": 1.734375, | |
| "eval_runtime": 5.8469, | |
| "eval_samples_per_second": 233.117, | |
| "eval_steps_per_second": 29.246, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "step": 25000, | |
| "total_flos": 8.355380415011226e+17, | |
| "train_loss": 1.84288763671875, | |
| "train_runtime": 4774.5297, | |
| "train_samples_per_second": 41.889, | |
| "train_steps_per_second": 5.236 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 2, | |
| "total_flos": 8.355380415011226e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |