trainer_state.json

{
  "best_metric": 0.9602272727272727,
  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-eurosat/checkpoint-36",
  "epoch": 2.88,
  "eval_steps": 500,
  "global_step": 36,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8,
      "grad_norm": 5.466047286987305,
      "learning_rate": 4.0625000000000005e-05,
      "loss": 0.6516,
      "step": 10
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.9318181818181818,
      "eval_loss": 0.36171892285346985,
      "eval_runtime": 1.4119,
      "eval_samples_per_second": 124.651,
      "eval_steps_per_second": 4.249,
      "step": 12
    },
    {
      "epoch": 1.6,
      "grad_norm": 8.815180778503418,
      "learning_rate": 2.5e-05,
      "loss": 0.3909,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9488636363636364,
      "eval_loss": 0.15098439157009125,
      "eval_runtime": 1.4258,
      "eval_samples_per_second": 123.442,
      "eval_steps_per_second": 4.208,
      "step": 25
    },
    {
      "epoch": 2.4,
      "grad_norm": 7.026189804077148,
      "learning_rate": 9.375000000000001e-06,
      "loss": 0.2274,
      "step": 30
    },
    {
      "epoch": 2.88,
      "eval_accuracy": 0.9602272727272727,
      "eval_loss": 0.09607520699501038,
      "eval_runtime": 1.5278,
      "eval_samples_per_second": 115.199,
      "eval_steps_per_second": 3.927,
      "step": 36
    },
    {
      "epoch": 2.88,
      "step": 36,
      "total_flos": 1.1334330977550336e+17,
      "train_loss": 0.3907758593559265,
      "train_runtime": 85.0457,
      "train_samples_per_second": 55.594,
      "train_steps_per_second": 0.423
    }
  ],
  "logging_steps": 10,
  "max_steps": 36,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.1334330977550336e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}