tmalsayed/hscode-classifier

1{2  "best_metric": null,3  "best_model_checkpoint": null,4  "epoch": 5.0,5  "eval_steps": 500,6  "global_step": 155,7  "is_hyper_param_search": false,8  "is_local_process_zero": true,9  "is_world_process_zero": true,10  "log_history": [11    {12      "epoch": 1.6129032258064515,13      "grad_norm": 0.660424530506134,14      "learning_rate": 3.387096774193548e-05,15      "loss": 9.0013,16      "step": 5017    },18    {19      "epoch": 3.225806451612903,20      "grad_norm": 0.7054967284202576,21      "learning_rate": 1.774193548387097e-05,22      "loss": 8.9754,23      "step": 10024    },25    {26      "epoch": 4.838709677419355,27      "grad_norm": 1.173000454902649,28      "learning_rate": 1.6129032258064516e-06,29      "loss": 8.9382,30      "step": 15031    }32  ],33  "logging_steps": 50,34  "max_steps": 155,35  "num_input_tokens_seen": 0,36  "num_train_epochs": 5,37  "save_steps": 500,38  "stateful_callbacks": {39    "TrainerControl": {40      "args": {41        "should_epoch_stop": false,42        "should_evaluate": false,43        "should_log": false,44        "should_save": true,45        "should_training_stop": true46      },47      "attributes": {}48    }49  },50  "total_flos": 1.11444326042112e+16,51  "train_batch_size": 256,52  "trial_name": null,53  "trial_params": null54}55