trainer_state.json

{
  "best_metric": 0.2957749664783478,
  "best_model_checkpoint": "pubmedbert_caption_compound_RTX6000/checkpoint-525",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 525,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 3.1260972023010254,
      "learning_rate": 2.4285714285714288e-05,
      "loss": 0.2901,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8874109263657958,
      "eval_f1": 0.8850925368018971,
      "eval_f1_medical": 0.8687707641196013,
      "eval_f1_non_medical": 0.901414309484193,
      "eval_loss": 0.2957749664783478,
      "eval_precision": 0.8820521098447567,
      "eval_recall": 0.891335859705761,
      "eval_runtime": 28.7719,
      "eval_samples_per_second": 146.323,
      "eval_steps_per_second": 4.588,
      "step": 525
    }
  ],
  "logging_steps": 500,
  "max_steps": 2625,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3739844096446080.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}