{
  "best_metric": 0.010909353382885456,
  "best_model_checkpoint": "emigomez/vit-cropped-faces/checkpoint-200",
  "epoch": 20.0,
  "eval_steps": 100,
  "global_step": 640,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3125,
      "grad_norm": 1.6401095390319824,
      "learning_rate": 0.000196875,
      "loss": 0.692,
      "step": 10
    },
    {
      "epoch": 0.625,
      "grad_norm": 1.1231616735458374,
      "learning_rate": 0.00019375000000000002,
      "loss": 0.2743,
      "step": 20
    },
    {
      "epoch": 0.9375,
      "grad_norm": 0.7256351709365845,
      "learning_rate": 0.000190625,
      "loss": 0.0795,
      "step": 30
    },
    {
      "epoch": 1.25,
      "grad_norm": 0.17406541109085083,
      "learning_rate": 0.0001875,
      "loss": 0.0414,
      "step": 40
    },
    {
      "epoch": 1.5625,
      "grad_norm": 6.439935684204102,
      "learning_rate": 0.000184375,
      "loss": 0.0465,
      "step": 50
    },
    {
      "epoch": 1.875,
      "grad_norm": 2.296980857849121,
      "learning_rate": 0.00018125000000000001,
      "loss": 0.0526,
      "step": 60
    },
    {
      "epoch": 2.1875,
      "grad_norm": 0.09786612540483475,
      "learning_rate": 0.000178125,
      "loss": 0.0372,
      "step": 70
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.07251214236021042,
      "learning_rate": 0.000175,
      "loss": 0.0158,
      "step": 80
    },
    {
      "epoch": 2.8125,
      "grad_norm": 0.06504912674427032,
      "learning_rate": 0.00017187500000000002,
      "loss": 0.1044,
      "step": 90
    },
    {
      "epoch": 3.125,
      "grad_norm": 0.08557004481554031,
      "learning_rate": 0.00016875,
      "loss": 0.0254,
      "step": 100
    },
    {
      "epoch": 3.125,
      "eval_accuracy": 1.0,
      "eval_loss": 0.013552392832934856,
      "eval_runtime": 1.5523,
      "eval_samples_per_second": 36.077,
      "eval_steps_per_second": 4.51,
      "step": 100
    },
    {
      "epoch": 3.4375,
      "grad_norm": 0.06062212213873863,
      "learning_rate": 0.000165625,
      "loss": 0.0122,
      "step": 110
    },
    {
      "epoch": 3.75,
      "grad_norm": 0.050220146775245667,
      "learning_rate": 0.00016250000000000002,
      "loss": 0.0098,
      "step": 120
    },
    {
      "epoch": 4.0625,
      "grad_norm": 0.04843802750110626,
      "learning_rate": 0.000159375,
      "loss": 0.0091,
      "step": 130
    },
    {
      "epoch": 4.375,
      "grad_norm": 0.04158018156886101,
      "learning_rate": 0.00015625,
      "loss": 0.0081,
      "step": 140
    },
    {
      "epoch": 4.6875,
      "grad_norm": 0.03899329528212547,
      "learning_rate": 0.000153125,
      "loss": 0.0075,
      "step": 150
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.05264412984251976,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.0069,
      "step": 160
    },
    {
      "epoch": 5.3125,
      "grad_norm": 0.036546722054481506,
      "learning_rate": 0.000146875,
      "loss": 0.0064,
      "step": 170
    },
    {
      "epoch": 5.625,
      "grad_norm": 0.035343725234270096,
      "learning_rate": 0.00014375,
      "loss": 0.006,
      "step": 180
    },
    {
      "epoch": 5.9375,
      "grad_norm": 0.03636905550956726,
      "learning_rate": 0.00014062500000000002,
      "loss": 0.0056,
      "step": 190
    },
    {
      "epoch": 6.25,
      "grad_norm": 0.033068690448999405,
      "learning_rate": 0.0001375,
      "loss": 0.0053,
      "step": 200
    },
    {
      "epoch": 6.25,
      "eval_accuracy": 1.0,
      "eval_loss": 0.010909353382885456,
      "eval_runtime": 1.5116,
      "eval_samples_per_second": 37.048,
      "eval_steps_per_second": 4.631,
      "step": 200
    },
    {
      "epoch": 6.5625,
      "grad_norm": 0.027593158185482025,
      "learning_rate": 0.000134375,
      "loss": 0.0049,
      "step": 210
    },
    {
      "epoch": 6.875,
      "grad_norm": 0.02486700937151909,
      "learning_rate": 0.00013125000000000002,
      "loss": 0.0047,
      "step": 220
    },
    {
      "epoch": 7.1875,
      "grad_norm": 0.025286702439188957,
      "learning_rate": 0.000128125,
      "loss": 0.0045,
      "step": 230
    },
    {
      "epoch": 7.5,
      "grad_norm": 0.02420409955084324,
      "learning_rate": 0.000125,
      "loss": 0.0042,
      "step": 240
    },
    {
      "epoch": 7.8125,
      "grad_norm": 0.02255989797413349,
      "learning_rate": 0.00012187500000000001,
      "loss": 0.0041,
      "step": 250
    },
    {
      "epoch": 8.125,
      "grad_norm": 0.02108701318502426,
      "learning_rate": 0.00011875,
      "loss": 0.0038,
      "step": 260
    },
    {
      "epoch": 8.4375,
      "grad_norm": 0.020098580047488213,
      "learning_rate": 0.000115625,
      "loss": 0.0037,
      "step": 270
    },
    {
      "epoch": 8.75,
      "grad_norm": 0.019474145025014877,
      "learning_rate": 0.00011250000000000001,
      "loss": 0.0036,
      "step": 280
    },
    {
      "epoch": 9.0625,
      "grad_norm": 0.018637843430042267,
      "learning_rate": 0.000109375,
      "loss": 0.0035,
      "step": 290
    },
    {
      "epoch": 9.375,
      "grad_norm": 0.018013661727309227,
      "learning_rate": 0.00010625000000000001,
      "loss": 0.0033,
      "step": 300
    },
    {
      "epoch": 9.375,
      "eval_accuracy": 1.0,
      "eval_loss": 0.013854571618139744,
      "eval_runtime": 1.54,
      "eval_samples_per_second": 36.363,
      "eval_steps_per_second": 4.545,
      "step": 300
    },
    {
      "epoch": 9.6875,
      "grad_norm": 0.01775149628520012,
      "learning_rate": 0.000103125,
      "loss": 0.0032,
      "step": 310
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.01715986803174019,
      "learning_rate": 0.0001,
      "loss": 0.0031,
      "step": 320
    },
    {
      "epoch": 10.3125,
      "grad_norm": 0.016615115106105804,
      "learning_rate": 9.687500000000001e-05,
      "loss": 0.003,
      "step": 330
    },
    {
      "epoch": 10.625,
      "grad_norm": 0.016151374205946922,
      "learning_rate": 9.375e-05,
      "loss": 0.0029,
      "step": 340
    },
    {
      "epoch": 10.9375,
      "grad_norm": 0.01536835078150034,
      "learning_rate": 9.062500000000001e-05,
      "loss": 0.0028,
      "step": 350
    },
    {
      "epoch": 11.25,
      "grad_norm": 0.014975355938076973,
      "learning_rate": 8.75e-05,
      "loss": 0.0027,
      "step": 360
    },
    {
      "epoch": 11.5625,
      "grad_norm": 0.014888996258378029,
      "learning_rate": 8.4375e-05,
      "loss": 0.0027,
      "step": 370
    },
    {
      "epoch": 11.875,
      "grad_norm": 0.014734145253896713,
      "learning_rate": 8.125000000000001e-05,
      "loss": 0.0026,
      "step": 380
    },
    {
      "epoch": 12.1875,
      "grad_norm": 0.014760036021471024,
      "learning_rate": 7.8125e-05,
      "loss": 0.0025,
      "step": 390
    },
    {
      "epoch": 12.5,
      "grad_norm": 0.013639590702950954,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.0025,
      "step": 400
    },
    {
      "epoch": 12.5,
      "eval_accuracy": 1.0,
      "eval_loss": 0.012760639190673828,
      "eval_runtime": 1.5727,
      "eval_samples_per_second": 35.608,
      "eval_steps_per_second": 4.451,
      "step": 400
    },
    {
      "epoch": 12.8125,
      "grad_norm": 0.013434410095214844,
      "learning_rate": 7.1875e-05,
      "loss": 0.0024,
      "step": 410
    },
    {
      "epoch": 13.125,
      "grad_norm": 0.013357667252421379,
      "learning_rate": 6.875e-05,
      "loss": 0.0024,
      "step": 420
    },
    {
      "epoch": 13.4375,
      "grad_norm": 0.01299892459064722,
      "learning_rate": 6.562500000000001e-05,
      "loss": 0.0023,
      "step": 430
    },
    {
      "epoch": 13.75,
      "grad_norm": 0.01252039521932602,
      "learning_rate": 6.25e-05,
      "loss": 0.0023,
      "step": 440
    },
    {
      "epoch": 14.0625,
      "grad_norm": 0.012350697070360184,
      "learning_rate": 5.9375e-05,
      "loss": 0.0022,
      "step": 450
    },
    {
      "epoch": 14.375,
      "grad_norm": 0.012504469603300095,
      "learning_rate": 5.6250000000000005e-05,
      "loss": 0.0022,
      "step": 460
    },
    {
      "epoch": 14.6875,
      "grad_norm": 0.01206807978451252,
      "learning_rate": 5.3125000000000004e-05,
      "loss": 0.0022,
      "step": 470
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.013769702985882759,
      "learning_rate": 5e-05,
      "loss": 0.0021,
      "step": 480
    },
    {
      "epoch": 15.3125,
      "grad_norm": 0.01167486421763897,
      "learning_rate": 4.6875e-05,
      "loss": 0.0021,
      "step": 490
    },
    {
      "epoch": 15.625,
      "grad_norm": 0.011528071016073227,
      "learning_rate": 4.375e-05,
      "loss": 0.0021,
      "step": 500
    },
    {
      "epoch": 15.625,
      "eval_accuracy": 1.0,
      "eval_loss": 0.012214643880724907,
      "eval_runtime": 1.5413,
      "eval_samples_per_second": 36.334,
      "eval_steps_per_second": 4.542,
      "step": 500
    },
    {
      "epoch": 15.9375,
      "grad_norm": 0.011460980400443077,
      "learning_rate": 4.0625000000000005e-05,
      "loss": 0.002,
      "step": 510
    },
    {
      "epoch": 16.25,
      "grad_norm": 0.011635087430477142,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.002,
      "step": 520
    },
    {
      "epoch": 16.5625,
      "grad_norm": 0.012200203724205494,
      "learning_rate": 3.4375e-05,
      "loss": 0.002,
      "step": 530
    },
    {
      "epoch": 16.875,
      "grad_norm": 0.011691519059240818,
      "learning_rate": 3.125e-05,
      "loss": 0.002,
      "step": 540
    },
    {
      "epoch": 17.1875,
      "grad_norm": 0.010916861705482006,
      "learning_rate": 2.8125000000000003e-05,
      "loss": 0.002,
      "step": 550
    },
    {
      "epoch": 17.5,
      "grad_norm": 0.010833012871444225,
      "learning_rate": 2.5e-05,
      "loss": 0.0019,
      "step": 560
    },
    {
      "epoch": 17.8125,
      "grad_norm": 0.010946971364319324,
      "learning_rate": 2.1875e-05,
      "loss": 0.0019,
      "step": 570
    },
    {
      "epoch": 18.125,
      "grad_norm": 0.01121476013213396,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.0019,
      "step": 580
    },
    {
      "epoch": 18.4375,
      "grad_norm": 0.010572544299066067,
      "learning_rate": 1.5625e-05,
      "loss": 0.0019,
      "step": 590
    },
    {
      "epoch": 18.75,
      "grad_norm": 0.01052548922598362,
      "learning_rate": 1.25e-05,
      "loss": 0.0019,
      "step": 600
    },
    {
      "epoch": 18.75,
      "eval_accuracy": 1.0,
      "eval_loss": 0.011984586715698242,
      "eval_runtime": 1.5235,
      "eval_samples_per_second": 36.758,
      "eval_steps_per_second": 4.595,
      "step": 600
    },
    {
      "epoch": 19.0625,
      "grad_norm": 0.011718451976776123,
      "learning_rate": 9.375000000000001e-06,
      "loss": 0.0019,
      "step": 610
    },
    {
      "epoch": 19.375,
      "grad_norm": 0.011368807405233383,
      "learning_rate": 6.25e-06,
      "loss": 0.0019,
      "step": 620
    },
    {
      "epoch": 19.6875,
      "grad_norm": 0.01071301568299532,
      "learning_rate": 3.125e-06,
      "loss": 0.0019,
      "step": 630
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.011912377551198006,
      "learning_rate": 0.0,
      "loss": 0.0019,
      "step": 640
    },
    {
      "epoch": 20.0,
      "step": 640,
      "total_flos": 7.764766951908557e+17,
      "train_loss": 0.024429786371183582,
      "train_runtime": 337.6815,
      "train_samples_per_second": 29.673,
      "train_steps_per_second": 1.895
    }
  ],
  "logging_steps": 10,
  "max_steps": 640,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 7.764766951908557e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}