| { | |
| "best_metric": 3.144331693649292, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_az_sft_1000samples_-1vocab_original-frozen/checkpoint-5000", | |
| "epoch": 144.9225225225225, | |
| "global_step": 20000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 18.12, | |
| "l1_reg_loss": 0.0, | |
| "learning_rate": 9e-05, | |
| "loss": 3.1058, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 36.23, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 8e-05, | |
| "loss": 2.3092, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 36.23, | |
| "eval_loss": 3.144331693649292, | |
| "eval_runtime": 516.01, | |
| "eval_samples_per_second": 9.508, | |
| "eval_steps_per_second": 4.754, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 54.35, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 7e-05, | |
| "loss": 1.854, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 72.46, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 6e-05, | |
| "loss": 1.4578, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 72.46, | |
| "eval_loss": 4.3338799476623535, | |
| "eval_runtime": 515.6849, | |
| "eval_samples_per_second": 9.514, | |
| "eval_steps_per_second": 4.757, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 90.58, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 5e-05, | |
| "loss": 1.121, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 108.69, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 4e-05, | |
| "loss": 0.8593, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 108.69, | |
| "eval_loss": 5.458686351776123, | |
| "eval_runtime": 514.7726, | |
| "eval_samples_per_second": 9.53, | |
| "eval_steps_per_second": 4.765, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 126.81, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 3e-05, | |
| "loss": 0.6664, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 144.92, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 2e-05, | |
| "loss": 0.5312, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 144.92, | |
| "eval_loss": 6.133657932281494, | |
| "eval_runtime": 514.9507, | |
| "eval_samples_per_second": 9.527, | |
| "eval_steps_per_second": 4.764, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 144.92, | |
| "step": 20000, | |
| "total_flos": 2.9879104118666035e+17, | |
| "train_loss": 1.488097003173828, | |
| "train_runtime": 49511.5153, | |
| "train_samples_per_second": 4.039, | |
| "train_steps_per_second": 0.505 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 182, | |
| "total_flos": 2.9879104118666035e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |