trainer_state.json

{
  "best_metric": 0.9494949494949495,
  "best_model_checkpoint": "mit-b0-finetuned-eurosat\\checkpoint-21",
  "epoch": 3.0,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8080808080808081,
      "eval_loss": 0.38276588916778564,
      "eval_runtime": 18.5626,
      "eval_samples_per_second": 5.333,
      "eval_steps_per_second": 0.215,
      "step": 7
    },
    {
      "epoch": 1.43,
      "learning_rate": 3.055555555555556e-05,
      "loss": 0.4864,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9191919191919192,
      "eval_loss": 0.2223997414112091,
      "eval_runtime": 18.4144,
      "eval_samples_per_second": 5.376,
      "eval_steps_per_second": 0.217,
      "step": 14
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.777777777777778e-06,
      "loss": 0.2035,
      "step": 20
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9494949494949495,
      "eval_loss": 0.17816942930221558,
      "eval_runtime": 17.3231,
      "eval_samples_per_second": 5.715,
      "eval_steps_per_second": 0.231,
      "step": 21
    },
    {
      "epoch": 3.0,
      "step": 21,
      "total_flos": 4.15443275785175e+16,
      "train_loss": 0.3380747139453888,
      "train_runtime": 1182.7003,
      "train_samples_per_second": 2.242,
      "train_steps_per_second": 0.018
    }
  ],
  "max_steps": 21,
  "num_train_epochs": 3,
  "total_flos": 4.15443275785175e+16,
  "trial_name": null,
  "trial_params": null
}