| { | |
| "best_metric": 2.728515625, | |
| "best_model_checkpoint": "/home/zhengxinyong/outputs/bloom-560m_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": 3.3833, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 0.0001, | |
| "loss": 2.9341, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_loss": 2.947265625, | |
| "eval_runtime": 3.039, | |
| "eval_samples_per_second": 264.56, | |
| "eval_steps_per_second": 33.235, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 0.0001, | |
| "loss": 2.7938, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 0.0001, | |
| "loss": 2.6616, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 2.814453125, | |
| "eval_runtime": 3.0893, | |
| "eval_samples_per_second": 260.252, | |
| "eval_steps_per_second": 32.693, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 0.0001, | |
| "loss": 2.5465, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0001, | |
| "loss": 2.5492, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_loss": 2.763671875, | |
| "eval_runtime": 3.0581, | |
| "eval_samples_per_second": 262.91, | |
| "eval_steps_per_second": 33.027, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 0.0001, | |
| "loss": 2.5363, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 0.0001, | |
| "loss": 2.3722, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 2.76171875, | |
| "eval_runtime": 3.0267, | |
| "eval_samples_per_second": 265.635, | |
| "eval_steps_per_second": 33.37, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 0.0001, | |
| "loss": 2.3422, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 0.0001, | |
| "loss": 2.3698, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "eval_loss": 2.728515625, | |
| "eval_runtime": 3.0354, | |
| "eval_samples_per_second": 264.873, | |
| "eval_steps_per_second": 33.274, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "step": 25000, | |
| "total_flos": 3.714828024466637e+17, | |
| "train_loss": 2.648919765625, | |
| "train_runtime": 4254.9078, | |
| "train_samples_per_second": 47.005, | |
| "train_steps_per_second": 5.876 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 3, | |
| "total_flos": 3.714828024466637e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |