| { | |
| "best_metric": 2.2680606842041016, | |
| "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_az_fish_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, | |
| "learning_rate": 9e-05, | |
| "loss": 2.9608, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 8e-05, | |
| "loss": 2.5218, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "eval_loss": 2.484114170074463, | |
| "eval_runtime": 513.9237, | |
| "eval_samples_per_second": 9.546, | |
| "eval_steps_per_second": 4.773, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.96, | |
| "learning_rate": 7e-05, | |
| "loss": 2.379, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "learning_rate": 6e-05, | |
| "loss": 2.2927, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "eval_loss": 2.3578550815582275, | |
| "eval_runtime": 514.0611, | |
| "eval_samples_per_second": 9.544, | |
| "eval_steps_per_second": 4.772, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 9.93, | |
| "learning_rate": 5e-05, | |
| "loss": 2.2308, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 11.91, | |
| "learning_rate": 4e-05, | |
| "loss": 2.1853, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 11.91, | |
| "eval_loss": 2.305732488632202, | |
| "eval_runtime": 512.0621, | |
| "eval_samples_per_second": 9.581, | |
| "eval_steps_per_second": 4.79, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 13.9, | |
| "learning_rate": 3e-05, | |
| "loss": 2.1504, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 15.89, | |
| "learning_rate": 2e-05, | |
| "loss": 2.1224, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 15.89, | |
| "eval_loss": 2.276737928390503, | |
| "eval_runtime": 511.8831, | |
| "eval_samples_per_second": 9.584, | |
| "eval_steps_per_second": 4.792, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 17.87, | |
| "learning_rate": 1e-05, | |
| "loss": 2.096, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "learning_rate": 0.0, | |
| "loss": 2.0827, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "eval_loss": 2.2680606842041016, | |
| "eval_runtime": 511.8308, | |
| "eval_samples_per_second": 9.585, | |
| "eval_steps_per_second": 4.793, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 19.86, | |
| "step": 25000, | |
| "total_flos": 3.7165924872093696e+17, | |
| "train_loss": 2.30218890625, | |
| "train_runtime": 60621.0247, | |
| "train_samples_per_second": 3.299, | |
| "train_steps_per_second": 0.412 | |
| } | |
| ], | |
| "max_steps": 25000, | |
| "num_train_epochs": 20, | |
| "total_flos": 3.7165924872093696e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |