| { | |
| "best_metric": 2.3144264221191406, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_az_sft_100000samples_-1vocab_original-frozen/checkpoint-25000", | |
| "epoch": 2.1258503401360542, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.21, | |
| "l1_reg_loss": 0.0, | |
| "learning_rate": 9e-05, | |
| "loss": 3.1768, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "l1_reg_loss": 0.0, | |
| "learning_rate": 8e-05, | |
| "loss": 2.6549, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "eval_loss": 2.602942943572998, | |
| "eval_runtime": 480.7936, | |
| "eval_samples_per_second": 9.584, | |
| "eval_steps_per_second": 4.792, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 7e-05, | |
| "loss": 2.5072, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 6e-05, | |
| "loss": 2.4336, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "eval_loss": 2.436041831970215, | |
| "eval_runtime": 480.6448, | |
| "eval_samples_per_second": 9.587, | |
| "eval_steps_per_second": 4.794, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 5e-05, | |
| "loss": 2.3747, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 4e-05, | |
| "loss": 2.332, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "eval_loss": 2.36257266998291, | |
| "eval_runtime": 480.5182, | |
| "eval_samples_per_second": 9.59, | |
| "eval_steps_per_second": 4.795, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 3e-05, | |
| "loss": 2.3094, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 2e-05, | |
| "loss": 2.2923, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "eval_loss": 2.326047658920288, | |
| "eval_runtime": 480.5697, | |
| "eval_samples_per_second": 9.589, | |
| "eval_steps_per_second": 4.794, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 1e-05, | |
| "loss": 2.2803, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 0.0, | |
| "loss": 2.2793, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_loss": 2.3144264221191406, | |
| "eval_runtime": 478.8598, | |
| "eval_samples_per_second": 9.623, | |
| "eval_steps_per_second": 4.811, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "step": 25000, | |
| "total_flos": 3.714827943936e+17, | |
| "train_loss": 2.46404990234375, | |
| "train_runtime": 61043.1273, | |
| "train_samples_per_second": 3.276, | |
| "train_steps_per_second": 0.41 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 3, | |
| "total_flos": 3.714827943936e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |