| { | |
| "best_metric": 2.3477182388305664, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_az_sft_10000samples_-1vocab_original-frozen/checkpoint-25000", | |
| "epoch": 19.856519150625125, | |
| "global_step": 25000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.99, | |
| "l1_reg_loss": 0.0, | |
| "learning_rate": 9e-05, | |
| "loss": 3.1946, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 8e-05, | |
| "loss": 2.6396, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "eval_loss": 2.5872137546539307, | |
| "eval_runtime": 514.4581, | |
| "eval_samples_per_second": 9.536, | |
| "eval_steps_per_second": 4.768, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.96, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 7e-05, | |
| "loss": 2.4815, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 6e-05, | |
| "loss": 2.3898, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "eval_loss": 2.439215898513794, | |
| "eval_runtime": 514.6109, | |
| "eval_samples_per_second": 9.533, | |
| "eval_steps_per_second": 4.767, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 9.93, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 5e-05, | |
| "loss": 2.3298, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 11.91, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 4e-05, | |
| "loss": 2.287, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 11.91, | |
| "eval_loss": 2.3796885013580322, | |
| "eval_runtime": 514.3961, | |
| "eval_samples_per_second": 9.537, | |
| "eval_steps_per_second": 4.769, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 13.9, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 3e-05, | |
| "loss": 2.254, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 15.89, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 2e-05, | |
| "loss": 2.2339, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 15.89, | |
| "eval_loss": 2.354508399963379, | |
| "eval_runtime": 514.4981, | |
| "eval_samples_per_second": 9.536, | |
| "eval_steps_per_second": 4.768, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 17.87, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 1e-05, | |
| "loss": 2.2184, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "l1_reg_loss": 0.0001, | |
| "learning_rate": 0.0, | |
| "loss": 2.21, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "eval_loss": 2.3477182388305664, | |
| "eval_runtime": 514.3805, | |
| "eval_samples_per_second": 9.538, | |
| "eval_steps_per_second": 4.769, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "step": 25000, | |
| "total_flos": 3.7165924872093696e+17, | |
| "train_loss": 2.42385498046875, | |
| "train_runtime": 61408.2472, | |
| "train_samples_per_second": 3.257, | |
| "train_steps_per_second": 0.407 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 20, | |
| "total_flos": 3.7165924872093696e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |