trainer_state.json

{
  "best_metric": 0.13999451001921492,
  "best_model_checkpoint": "ds3-img-classification/checkpoint-35",
  "epoch": 0.9922480620155039,
  "eval_steps": 35,
  "global_step": 64,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.015503875968992248,
      "grad_norm": 12.836527824401855,
      "learning_rate": 0.0002857142857142857,
      "loss": 5.7815,
      "step": 1
    },
    {
      "epoch": 0.5426356589147286,
      "grad_norm": 6.243555545806885,
      "learning_rate": 0.0010275543423681622,
      "loss": 4.9108,
      "step": 35
    },
    {
      "epoch": 0.5426356589147286,
      "eval_accuracy": 0.13999451001921492,
      "eval_loss": 5.21905517578125,
      "eval_macro_f1": 0.04648934733573823,
      "eval_macro_precision": 0.03716366453858833,
      "eval_macro_recall": 0.09867228278248098,
      "eval_micro_f1": 0.13999451001921492,
      "eval_micro_precision": 0.13999451001921492,
      "eval_micro_recall": 0.13999451001921492,
      "eval_runtime": 8.2153,
      "eval_samples_per_second": 443.44,
      "eval_steps_per_second": 3.53,
      "eval_weighted_f1": 0.0724726180615966,
      "eval_weighted_precision": 0.06178867753876749,
      "eval_weighted_recall": 0.13999451001921492,
      "step": 35
    },
    {
      "epoch": 0.9922480620155039,
      "step": 64,
      "total_flos": 8.978215898519175e+18,
      "train_loss": 4.277425870299339,
      "train_runtime": 163.0123,
      "train_samples_per_second": 201.126,
      "train_steps_per_second": 0.393
    }
  ],
  "logging_steps": 35,
  "max_steps": 64,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 35,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.978215898519175e+18,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}