{
  "best_metric": 0.8693982074263764,
  "best_model_checkpoint": "swin-tiny-patch4-window7-224-MM_Classification/checkpoint-361",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 380,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5263157894736842,
      "grad_norm": 3.3412868976593018,
      "learning_rate": 1.3157894736842106e-05,
      "loss": 1.0476,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6530089628681178,
      "eval_loss": 0.7707358002662659,
      "eval_runtime": 104.799,
      "eval_samples_per_second": 7.452,
      "eval_steps_per_second": 0.067,
      "step": 19
    },
    {
      "epoch": 1.0526315789473684,
      "grad_norm": 3.460216999053955,
      "learning_rate": 2.6315789473684212e-05,
      "loss": 0.7869,
      "step": 20
    },
    {
      "epoch": 1.5789473684210527,
      "grad_norm": 3.0562336444854736,
      "learning_rate": 3.9473684210526316e-05,
      "loss": 0.6226,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8104993597951344,
      "eval_loss": 0.47430700063705444,
      "eval_runtime": 75.387,
      "eval_samples_per_second": 10.36,
      "eval_steps_per_second": 0.093,
      "step": 38
    },
    {
      "epoch": 2.1052631578947367,
      "grad_norm": 3.6400160789489746,
      "learning_rate": 4.970760233918128e-05,
      "loss": 0.5102,
      "step": 40
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 2.88069486618042,
      "learning_rate": 4.824561403508772e-05,
      "loss": 0.4477,
      "step": 50
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8322663252240717,
      "eval_loss": 0.41332316398620605,
      "eval_runtime": 75.6871,
      "eval_samples_per_second": 10.319,
      "eval_steps_per_second": 0.092,
      "step": 57
    },
    {
      "epoch": 3.1578947368421053,
      "grad_norm": 2.5295557975769043,
      "learning_rate": 4.678362573099415e-05,
      "loss": 0.4205,
      "step": 60
    },
    {
      "epoch": 3.6842105263157894,
      "grad_norm": 2.7619762420654297,
      "learning_rate": 4.5321637426900585e-05,
      "loss": 0.3963,
      "step": 70
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8476312419974392,
      "eval_loss": 0.38133054971694946,
      "eval_runtime": 75.5986,
      "eval_samples_per_second": 10.331,
      "eval_steps_per_second": 0.093,
      "step": 76
    },
    {
      "epoch": 4.2105263157894735,
      "grad_norm": 2.971311330795288,
      "learning_rate": 4.3859649122807014e-05,
      "loss": 0.3788,
      "step": 80
    },
    {
      "epoch": 4.7368421052631575,
      "grad_norm": 3.20125412940979,
      "learning_rate": 4.239766081871345e-05,
      "loss": 0.3694,
      "step": 90
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8540332906530089,
      "eval_loss": 0.37533658742904663,
      "eval_runtime": 74.9213,
      "eval_samples_per_second": 10.424,
      "eval_steps_per_second": 0.093,
      "step": 95
    },
    {
      "epoch": 5.2631578947368425,
      "grad_norm": 3.091187000274658,
      "learning_rate": 4.093567251461988e-05,
      "loss": 0.3499,
      "step": 100
    },
    {
      "epoch": 5.7894736842105265,
      "grad_norm": 3.088123321533203,
      "learning_rate": 3.9473684210526316e-05,
      "loss": 0.3451,
      "step": 110
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8489116517285531,
      "eval_loss": 0.3586506247520447,
      "eval_runtime": 75.7408,
      "eval_samples_per_second": 10.311,
      "eval_steps_per_second": 0.092,
      "step": 114
    },
    {
      "epoch": 6.315789473684211,
      "grad_norm": 2.5398190021514893,
      "learning_rate": 3.8011695906432746e-05,
      "loss": 0.3421,
      "step": 120
    },
    {
      "epoch": 6.842105263157895,
      "grad_norm": 3.2862489223480225,
      "learning_rate": 3.654970760233918e-05,
      "loss": 0.3382,
      "step": 130
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8450704225352113,
      "eval_loss": 0.3531467616558075,
      "eval_runtime": 75.2863,
      "eval_samples_per_second": 10.374,
      "eval_steps_per_second": 0.093,
      "step": 133
    },
    {
      "epoch": 7.368421052631579,
      "grad_norm": 2.175835609436035,
      "learning_rate": 3.508771929824561e-05,
      "loss": 0.3397,
      "step": 140
    },
    {
      "epoch": 7.894736842105263,
      "grad_norm": 2.2023236751556396,
      "learning_rate": 3.362573099415205e-05,
      "loss": 0.3253,
      "step": 150
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8578745198463509,
      "eval_loss": 0.34979528188705444,
      "eval_runtime": 75.6534,
      "eval_samples_per_second": 10.323,
      "eval_steps_per_second": 0.093,
      "step": 152
    },
    {
      "epoch": 8.421052631578947,
      "grad_norm": 3.327239513397217,
      "learning_rate": 3.216374269005848e-05,
      "loss": 0.3156,
      "step": 160
    },
    {
      "epoch": 8.947368421052632,
      "grad_norm": 3.0708658695220947,
      "learning_rate": 3.0701754385964913e-05,
      "loss": 0.3121,
      "step": 170
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8578745198463509,
      "eval_loss": 0.34373539686203003,
      "eval_runtime": 74.9284,
      "eval_samples_per_second": 10.423,
      "eval_steps_per_second": 0.093,
      "step": 171
    },
    {
      "epoch": 9.473684210526315,
      "grad_norm": 2.4968678951263428,
      "learning_rate": 2.9239766081871346e-05,
      "loss": 0.3042,
      "step": 180
    },
    {
      "epoch": 10.0,
      "grad_norm": 2.865316152572632,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.2855,
      "step": 190
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8655569782330346,
      "eval_loss": 0.3446912467479706,
      "eval_runtime": 75.4844,
      "eval_samples_per_second": 10.347,
      "eval_steps_per_second": 0.093,
      "step": 190
    },
    {
      "epoch": 10.526315789473685,
      "grad_norm": 3.000030279159546,
      "learning_rate": 2.6315789473684212e-05,
      "loss": 0.2961,
      "step": 200
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8617157490396927,
      "eval_loss": 0.3350251019001007,
      "eval_runtime": 74.9625,
      "eval_samples_per_second": 10.419,
      "eval_steps_per_second": 0.093,
      "step": 209
    },
    {
      "epoch": 11.052631578947368,
      "grad_norm": 3.1556169986724854,
      "learning_rate": 2.485380116959064e-05,
      "loss": 0.291,
      "step": 210
    },
    {
      "epoch": 11.578947368421053,
      "grad_norm": 2.82590389251709,
      "learning_rate": 2.3391812865497074e-05,
      "loss": 0.273,
      "step": 220
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8565941101152369,
      "eval_loss": 0.34841132164001465,
      "eval_runtime": 75.6499,
      "eval_samples_per_second": 10.324,
      "eval_steps_per_second": 0.093,
      "step": 228
    },
    {
      "epoch": 12.105263157894736,
      "grad_norm": 2.110739231109619,
      "learning_rate": 2.1929824561403507e-05,
      "loss": 0.2767,
      "step": 230
    },
    {
      "epoch": 12.631578947368421,
      "grad_norm": 2.7739641666412354,
      "learning_rate": 2.046783625730994e-05,
      "loss": 0.2745,
      "step": 240
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8604353393085787,
      "eval_loss": 0.34332236647605896,
      "eval_runtime": 75.3714,
      "eval_samples_per_second": 10.362,
      "eval_steps_per_second": 0.093,
      "step": 247
    },
    {
      "epoch": 13.157894736842104,
      "grad_norm": 2.834440231323242,
      "learning_rate": 1.9005847953216373e-05,
      "loss": 0.2678,
      "step": 250
    },
    {
      "epoch": 13.68421052631579,
      "grad_norm": 4.303690433502197,
      "learning_rate": 1.7543859649122806e-05,
      "loss": 0.2613,
      "step": 260
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8642765685019206,
      "eval_loss": 0.34982678294181824,
      "eval_runtime": 76.0549,
      "eval_samples_per_second": 10.269,
      "eval_steps_per_second": 0.092,
      "step": 266
    },
    {
      "epoch": 14.210526315789474,
      "grad_norm": 3.9326910972595215,
      "learning_rate": 1.608187134502924e-05,
      "loss": 0.2713,
      "step": 270
    },
    {
      "epoch": 14.736842105263158,
      "grad_norm": 3.0511579513549805,
      "learning_rate": 1.4619883040935673e-05,
      "loss": 0.2527,
      "step": 280
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8578745198463509,
      "eval_loss": 0.33652085065841675,
      "eval_runtime": 78.4572,
      "eval_samples_per_second": 9.954,
      "eval_steps_per_second": 0.089,
      "step": 285
    },
    {
      "epoch": 15.263157894736842,
      "grad_norm": 2.9660778045654297,
      "learning_rate": 1.3157894736842106e-05,
      "loss": 0.246,
      "step": 290
    },
    {
      "epoch": 15.789473684210526,
      "grad_norm": 2.621548652648926,
      "learning_rate": 1.1695906432748537e-05,
      "loss": 0.2619,
      "step": 300
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8617157490396927,
      "eval_loss": 0.3450033366680145,
      "eval_runtime": 78.025,
      "eval_samples_per_second": 10.01,
      "eval_steps_per_second": 0.09,
      "step": 304
    },
    {
      "epoch": 16.31578947368421,
      "grad_norm": 2.7999181747436523,
      "learning_rate": 1.023391812865497e-05,
      "loss": 0.2469,
      "step": 310
    },
    {
      "epoch": 16.842105263157894,
      "grad_norm": 2.3670365810394287,
      "learning_rate": 8.771929824561403e-06,
      "loss": 0.2436,
      "step": 320
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8681177976952625,
      "eval_loss": 0.34535887837409973,
      "eval_runtime": 78.3498,
      "eval_samples_per_second": 9.968,
      "eval_steps_per_second": 0.089,
      "step": 323
    },
    {
      "epoch": 17.36842105263158,
      "grad_norm": 2.8293299674987793,
      "learning_rate": 7.3099415204678366e-06,
      "loss": 0.2457,
      "step": 330
    },
    {
      "epoch": 17.894736842105264,
      "grad_norm": 2.825676441192627,
      "learning_rate": 5.8479532163742686e-06,
      "loss": 0.2518,
      "step": 340
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8681177976952625,
      "eval_loss": 0.3437488377094269,
      "eval_runtime": 77.0059,
      "eval_samples_per_second": 10.142,
      "eval_steps_per_second": 0.091,
      "step": 342
    },
    {
      "epoch": 18.42105263157895,
      "grad_norm": 2.9020776748657227,
      "learning_rate": 4.3859649122807014e-06,
      "loss": 0.2362,
      "step": 350
    },
    {
      "epoch": 18.94736842105263,
      "grad_norm": 2.6419830322265625,
      "learning_rate": 2.9239766081871343e-06,
      "loss": 0.243,
      "step": 360
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8693982074263764,
      "eval_loss": 0.34680071473121643,
      "eval_runtime": 76.7708,
      "eval_samples_per_second": 10.173,
      "eval_steps_per_second": 0.091,
      "step": 361
    },
    {
      "epoch": 19.473684210526315,
      "grad_norm": 2.185123920440674,
      "learning_rate": 1.4619883040935671e-06,
      "loss": 0.2452,
      "step": 370
    },
    {
      "epoch": 20.0,
      "grad_norm": 4.370210647583008,
      "learning_rate": 0.0,
      "loss": 0.2415,
      "step": 380
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8693982074263764,
      "eval_loss": 0.3454751670360565,
      "eval_runtime": 76.6964,
      "eval_samples_per_second": 10.183,
      "eval_steps_per_second": 0.091,
      "step": 380
    },
    {
      "epoch": 20.0,
      "step": 380,
      "total_flos": 4.783917310653358e+18,
      "train_loss": 0.3464228366550646,
      "train_runtime": 17694.0061,
      "train_samples_per_second": 10.877,
      "train_steps_per_second": 0.021
    }
  ],
  "logging_steps": 10,
  "max_steps": 380,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "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": 4.783917310653358e+18,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}