trainer_state.json

{
  "best_metric": 3.3893251419067383,
  "best_model_checkpoint": "./qwen2vl_medical_finetuned_from_step4/checkpoint-45",
  "epoch": 2.845528455284553,
  "eval_steps": 500,
  "global_step": 45,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 3.4156579971313477,
      "eval_mean_token_accuracy": 0.397203758839638,
      "eval_num_tokens": 102536.0,
      "eval_runtime": 241.6603,
      "eval_samples_per_second": 1.014,
      "eval_steps_per_second": 0.257,
      "step": 16
    },
    {
      "epoch": 1.2601626016260163,
      "grad_norm": 9.601910591125488,
      "learning_rate": 5.555555555555557e-06,
      "loss": 3.3201,
      "mean_token_accuracy": 0.3937227354895684,
      "num_tokens": 129456.0,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_loss": 3.395120143890381,
      "eval_mean_token_accuracy": 0.3980548819226603,
      "eval_num_tokens": 205072.0,
      "eval_runtime": 241.3279,
      "eval_samples_per_second": 1.015,
      "eval_steps_per_second": 0.257,
      "step": 32
    },
    {
      "epoch": 2.5203252032520327,
      "grad_norm": 10.301806449890137,
      "learning_rate": 1.111111111111111e-06,
      "loss": 3.3005,
      "mean_token_accuracy": 0.3955008022246822,
      "num_tokens": 258653.0,
      "step": 40
    },
    {
      "epoch": 2.845528455284553,
      "eval_loss": 3.3893251419067383,
      "eval_mean_token_accuracy": 0.3977832736507539,
      "eval_num_tokens": 292154.0,
      "eval_runtime": 241.3615,
      "eval_samples_per_second": 1.015,
      "eval_steps_per_second": 0.257,
      "step": 45
    },
    {
      "epoch": 2.845528455284553,
      "mean_token_accuracy": 0.3963217779994011,
      "num_tokens": 292154.0,
      "step": 45,
      "total_flos": 3541381099060224.0,
      "train_loss": 3.3185320536295575,
      "train_runtime": 1691.5178,
      "train_samples_per_second": 0.435,
      "train_steps_per_second": 0.027
    }
  ],
  "logging_steps": 20,
  "max_steps": 45,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3541381099060224.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}