| { | |
| "best_metric": 2.947911262512207, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-560m_de_continual-pretrain_100000samples_-1vocab_original_bsz1/checkpoint-25000", | |
| "epoch": 0.34345377112240694, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 9e-05, | |
| "loss": 4.0829, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 8e-05, | |
| "loss": 3.6084, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "eval_loss": 3.581524133682251, | |
| "eval_runtime": 84.6587, | |
| "eval_samples_per_second": 9.497, | |
| "eval_steps_per_second": 4.748, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 7e-05, | |
| "loss": 3.4193, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 6e-05, | |
| "loss": 3.3097, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "eval_loss": 3.3223936557769775, | |
| "eval_runtime": 84.6499, | |
| "eval_samples_per_second": 9.498, | |
| "eval_steps_per_second": 4.749, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 5e-05, | |
| "loss": 3.2085, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4e-05, | |
| "loss": 3.0969, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_loss": 3.157708168029785, | |
| "eval_runtime": 84.5971, | |
| "eval_samples_per_second": 9.504, | |
| "eval_steps_per_second": 4.752, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 3e-05, | |
| "loss": 3.0205, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 2e-05, | |
| "loss": 2.968, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "eval_loss": 3.02886700630188, | |
| "eval_runtime": 84.6789, | |
| "eval_samples_per_second": 9.495, | |
| "eval_steps_per_second": 4.747, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 1e-05, | |
| "loss": 2.8963, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 0.0, | |
| "loss": 2.8626, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_loss": 2.947911262512207, | |
| "eval_runtime": 84.657, | |
| "eval_samples_per_second": 9.497, | |
| "eval_steps_per_second": 4.749, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "step": 25000, | |
| "total_flos": 4.64353492992e+16, | |
| "train_loss": 3.24729939453125, | |
| "train_runtime": 10361.3203, | |
| "train_samples_per_second": 2.413, | |
| "train_steps_per_second": 2.413 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 1, | |
| "total_flos": 4.64353492992e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |