trainer_state.json

{
  "best_metric": 0.8,
  "best_model_checkpoint": "vit-base-patch16-224/checkpoint-8",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.5545626282691956,
      "eval_runtime": 0.0989,
      "eval_samples_per_second": 151.64,
      "eval_steps_per_second": 40.437,
      "step": 8
    },
    {
      "epoch": 1.25,
      "grad_norm": 4.057127952575684,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.5945,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.5408901572227478,
      "eval_runtime": 0.1082,
      "eval_samples_per_second": 138.616,
      "eval_steps_per_second": 36.964,
      "step": 16
    },
    {
      "epoch": 2.5,
      "grad_norm": 4.53317403793335,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.5832,
      "step": 20
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.5466590523719788,
      "eval_runtime": 0.1022,
      "eval_samples_per_second": 146.703,
      "eval_steps_per_second": 39.121,
      "step": 24
    },
    {
      "epoch": 3.75,
      "grad_norm": 3.683661937713623,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.5338,
      "step": 30
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.551804780960083,
      "eval_runtime": 0.0956,
      "eval_samples_per_second": 156.964,
      "eval_steps_per_second": 41.857,
      "step": 32
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.659008979797363,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.5513,
      "step": 40
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.5602456331253052,
      "eval_runtime": 0.0943,
      "eval_samples_per_second": 159.134,
      "eval_steps_per_second": 42.436,
      "step": 40
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5607239007949829,
      "eval_runtime": 0.1057,
      "eval_samples_per_second": 141.965,
      "eval_steps_per_second": 37.857,
      "step": 48
    },
    {
      "epoch": 6.25,
      "grad_norm": 6.287199974060059,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.5417,
      "step": 50
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5707207322120667,
      "eval_runtime": 0.1098,
      "eval_samples_per_second": 136.606,
      "eval_steps_per_second": 36.428,
      "step": 56
    },
    {
      "epoch": 7.5,
      "grad_norm": 3.7512075901031494,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.5343,
      "step": 60
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5747508406639099,
      "eval_runtime": 0.0961,
      "eval_samples_per_second": 156.038,
      "eval_steps_per_second": 41.61,
      "step": 64
    },
    {
      "epoch": 8.75,
      "grad_norm": 4.6297926902771,
      "learning_rate": 6.944444444444446e-07,
      "loss": 0.5379,
      "step": 70
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5735542178153992,
      "eval_runtime": 0.0985,
      "eval_samples_per_second": 152.245,
      "eval_steps_per_second": 40.599,
      "step": 72
    },
    {
      "epoch": 10.0,
      "grad_norm": 2.6508984565734863,
      "learning_rate": 0.0,
      "loss": 0.5137,
      "step": 80
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.5730186700820923,
      "eval_runtime": 0.344,
      "eval_samples_per_second": 43.601,
      "eval_steps_per_second": 11.627,
      "step": 80
    },
    {
      "epoch": 10.0,
      "step": 80,
      "total_flos": 9.841482681053184e+16,
      "train_loss": 0.5487962663173676,
      "train_runtime": 36.3019,
      "train_samples_per_second": 34.984,
      "train_steps_per_second": 2.204
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "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": 9.841482681053184e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}