{
  "best_metric": 0.8333333333333334,
  "best_model_checkpoint": "vit-base-patch16-224-dmae-va-U5-42/checkpoint-108",
  "epoch": 37.935483870967744,
  "eval_steps": 500,
  "global_step": 294,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9,
      "eval_accuracy": 0.5,
      "eval_loss": 1.28582763671875,
      "eval_runtime": 1.6082,
      "eval_samples_per_second": 37.31,
      "eval_steps_per_second": 1.244,
      "step": 7
    },
    {
      "epoch": 1.55,
      "grad_norm": 4.701062202453613,
      "learning_rate": 2e-05,
      "loss": 1.3455,
      "step": 12
    },
    {
      "epoch": 1.94,
      "eval_accuracy": 0.48333333333333334,
      "eval_loss": 1.1091300249099731,
      "eval_runtime": 2.0547,
      "eval_samples_per_second": 29.201,
      "eval_steps_per_second": 0.973,
      "step": 15
    },
    {
      "epoch": 2.97,
      "eval_accuracy": 0.5833333333333334,
      "eval_loss": 0.8518259525299072,
      "eval_runtime": 1.5248,
      "eval_samples_per_second": 39.349,
      "eval_steps_per_second": 1.312,
      "step": 23
    },
    {
      "epoch": 3.1,
      "grad_norm": 2.8079495429992676,
      "learning_rate": 4e-05,
      "loss": 1.0067,
      "step": 24
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7166666666666667,
      "eval_loss": 0.7317402362823486,
      "eval_runtime": 1.5389,
      "eval_samples_per_second": 38.988,
      "eval_steps_per_second": 1.3,
      "step": 31
    },
    {
      "epoch": 4.65,
      "grad_norm": 3.14342999458313,
      "learning_rate": 4.886363636363637e-05,
      "loss": 0.6085,
      "step": 36
    },
    {
      "epoch": 4.9,
      "eval_accuracy": 0.75,
      "eval_loss": 0.694876492023468,
      "eval_runtime": 1.5371,
      "eval_samples_per_second": 39.035,
      "eval_steps_per_second": 1.301,
      "step": 38
    },
    {
      "epoch": 5.94,
      "eval_accuracy": 0.75,
      "eval_loss": 0.6633104085922241,
      "eval_runtime": 1.5303,
      "eval_samples_per_second": 39.208,
      "eval_steps_per_second": 1.307,
      "step": 46
    },
    {
      "epoch": 6.19,
      "grad_norm": 2.108163833618164,
      "learning_rate": 4.659090909090909e-05,
      "loss": 0.3389,
      "step": 48
    },
    {
      "epoch": 6.97,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.6791020035743713,
      "eval_runtime": 1.5402,
      "eval_samples_per_second": 38.956,
      "eval_steps_per_second": 1.299,
      "step": 54
    },
    {
      "epoch": 7.74,
      "grad_norm": 1.7717548608779907,
      "learning_rate": 4.431818181818182e-05,
      "loss": 0.1977,
      "step": 60
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.7010270357131958,
      "eval_runtime": 1.6107,
      "eval_samples_per_second": 37.252,
      "eval_steps_per_second": 1.242,
      "step": 62
    },
    {
      "epoch": 8.9,
      "eval_accuracy": 0.75,
      "eval_loss": 0.697003960609436,
      "eval_runtime": 1.5576,
      "eval_samples_per_second": 38.522,
      "eval_steps_per_second": 1.284,
      "step": 69
    },
    {
      "epoch": 9.29,
      "grad_norm": 1.6967554092407227,
      "learning_rate": 4.204545454545455e-05,
      "loss": 0.1496,
      "step": 72
    },
    {
      "epoch": 9.94,
      "eval_accuracy": 0.8,
      "eval_loss": 0.6983956098556519,
      "eval_runtime": 1.5413,
      "eval_samples_per_second": 38.929,
      "eval_steps_per_second": 1.298,
      "step": 77
    },
    {
      "epoch": 10.84,
      "grad_norm": 2.1696012020111084,
      "learning_rate": 3.9772727272727275e-05,
      "loss": 0.1194,
      "step": 84
    },
    {
      "epoch": 10.97,
      "eval_accuracy": 0.7333333333333333,
      "eval_loss": 0.9060508012771606,
      "eval_runtime": 1.5628,
      "eval_samples_per_second": 38.392,
      "eval_steps_per_second": 1.28,
      "step": 85
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.75,
      "eval_loss": 0.8720260858535767,
      "eval_runtime": 1.5352,
      "eval_samples_per_second": 39.084,
      "eval_steps_per_second": 1.303,
      "step": 93
    },
    {
      "epoch": 12.39,
      "grad_norm": 1.1283667087554932,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.109,
      "step": 96
    },
    {
      "epoch": 12.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.8438921570777893,
      "eval_runtime": 1.549,
      "eval_samples_per_second": 38.735,
      "eval_steps_per_second": 1.291,
      "step": 100
    },
    {
      "epoch": 13.94,
      "grad_norm": 1.2174146175384521,
      "learning_rate": 3.522727272727273e-05,
      "loss": 0.0902,
      "step": 108
    },
    {
      "epoch": 13.94,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.7344614863395691,
      "eval_runtime": 1.5138,
      "eval_samples_per_second": 39.635,
      "eval_steps_per_second": 1.321,
      "step": 108
    },
    {
      "epoch": 14.97,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.8419939279556274,
      "eval_runtime": 2.094,
      "eval_samples_per_second": 28.654,
      "eval_steps_per_second": 0.955,
      "step": 116
    },
    {
      "epoch": 15.48,
      "grad_norm": 2.3597609996795654,
      "learning_rate": 3.295454545454545e-05,
      "loss": 0.0938,
      "step": 120
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.75,
      "eval_loss": 0.799355149269104,
      "eval_runtime": 1.6114,
      "eval_samples_per_second": 37.235,
      "eval_steps_per_second": 1.241,
      "step": 124
    },
    {
      "epoch": 16.9,
      "eval_accuracy": 0.8,
      "eval_loss": 0.834147036075592,
      "eval_runtime": 1.5793,
      "eval_samples_per_second": 37.991,
      "eval_steps_per_second": 1.266,
      "step": 131
    },
    {
      "epoch": 17.03,
      "grad_norm": 3.3779869079589844,
      "learning_rate": 3.068181818181818e-05,
      "loss": 0.0862,
      "step": 132
    },
    {
      "epoch": 17.94,
      "eval_accuracy": 0.8,
      "eval_loss": 0.7238650918006897,
      "eval_runtime": 1.5475,
      "eval_samples_per_second": 38.771,
      "eval_steps_per_second": 1.292,
      "step": 139
    },
    {
      "epoch": 18.58,
      "grad_norm": 1.2446733713150024,
      "learning_rate": 2.8409090909090912e-05,
      "loss": 0.0864,
      "step": 144
    },
    {
      "epoch": 18.97,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.8485052585601807,
      "eval_runtime": 1.5613,
      "eval_samples_per_second": 38.428,
      "eval_steps_per_second": 1.281,
      "step": 147
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8,
      "eval_loss": 0.8948166370391846,
      "eval_runtime": 1.558,
      "eval_samples_per_second": 38.512,
      "eval_steps_per_second": 1.284,
      "step": 155
    },
    {
      "epoch": 20.13,
      "grad_norm": 1.548230767250061,
      "learning_rate": 2.6136363636363637e-05,
      "loss": 0.065,
      "step": 156
    },
    {
      "epoch": 20.9,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.8681192994117737,
      "eval_runtime": 1.5486,
      "eval_samples_per_second": 38.745,
      "eval_steps_per_second": 1.292,
      "step": 162
    },
    {
      "epoch": 21.68,
      "grad_norm": 1.3194923400878906,
      "learning_rate": 2.3863636363636365e-05,
      "loss": 0.0793,
      "step": 168
    },
    {
      "epoch": 21.94,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.8225926756858826,
      "eval_runtime": 1.5256,
      "eval_samples_per_second": 39.329,
      "eval_steps_per_second": 1.311,
      "step": 170
    },
    {
      "epoch": 22.97,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.7495377063751221,
      "eval_runtime": 1.5935,
      "eval_samples_per_second": 37.653,
      "eval_steps_per_second": 1.255,
      "step": 178
    },
    {
      "epoch": 23.23,
      "grad_norm": 1.1563166379928589,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.0629,
      "step": 180
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.7666666666666667,
      "eval_loss": 0.8813876509666443,
      "eval_runtime": 1.5206,
      "eval_samples_per_second": 39.458,
      "eval_steps_per_second": 1.315,
      "step": 186
    },
    {
      "epoch": 24.77,
      "grad_norm": 1.163840413093567,
      "learning_rate": 1.9318181818181818e-05,
      "loss": 0.0666,
      "step": 192
    },
    {
      "epoch": 24.9,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.7739368081092834,
      "eval_runtime": 1.5452,
      "eval_samples_per_second": 38.83,
      "eval_steps_per_second": 1.294,
      "step": 193
    },
    {
      "epoch": 25.94,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9246408939361572,
      "eval_runtime": 1.5689,
      "eval_samples_per_second": 38.244,
      "eval_steps_per_second": 1.275,
      "step": 201
    },
    {
      "epoch": 26.32,
      "grad_norm": 0.9550792574882507,
      "learning_rate": 1.7045454545454546e-05,
      "loss": 0.0571,
      "step": 204
    },
    {
      "epoch": 26.97,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.8077472448348999,
      "eval_runtime": 1.5388,
      "eval_samples_per_second": 38.992,
      "eval_steps_per_second": 1.3,
      "step": 209
    },
    {
      "epoch": 27.87,
      "grad_norm": 1.725164532661438,
      "learning_rate": 1.4772727272727274e-05,
      "loss": 0.0519,
      "step": 216
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.8974602818489075,
      "eval_runtime": 1.5489,
      "eval_samples_per_second": 38.738,
      "eval_steps_per_second": 1.291,
      "step": 217
    },
    {
      "epoch": 28.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9199196100234985,
      "eval_runtime": 1.5274,
      "eval_samples_per_second": 39.283,
      "eval_steps_per_second": 1.309,
      "step": 224
    },
    {
      "epoch": 29.42,
      "grad_norm": 0.8022506833076477,
      "learning_rate": 1.25e-05,
      "loss": 0.0523,
      "step": 228
    },
    {
      "epoch": 29.94,
      "eval_accuracy": 0.8,
      "eval_loss": 0.8512372374534607,
      "eval_runtime": 1.5327,
      "eval_samples_per_second": 39.145,
      "eval_steps_per_second": 1.305,
      "step": 232
    },
    {
      "epoch": 30.97,
      "grad_norm": 1.4496326446533203,
      "learning_rate": 1.0227272727272729e-05,
      "loss": 0.0548,
      "step": 240
    },
    {
      "epoch": 30.97,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.9376980066299438,
      "eval_runtime": 1.5279,
      "eval_samples_per_second": 39.268,
      "eval_steps_per_second": 1.309,
      "step": 240
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.8212655782699585,
      "eval_runtime": 1.5365,
      "eval_samples_per_second": 39.049,
      "eval_steps_per_second": 1.302,
      "step": 248
    },
    {
      "epoch": 32.52,
      "grad_norm": 2.0979461669921875,
      "learning_rate": 7.954545454545455e-06,
      "loss": 0.0576,
      "step": 252
    },
    {
      "epoch": 32.9,
      "eval_accuracy": 0.8166666666666667,
      "eval_loss": 0.8384222388267517,
      "eval_runtime": 1.5251,
      "eval_samples_per_second": 39.342,
      "eval_steps_per_second": 1.311,
      "step": 255
    },
    {
      "epoch": 33.94,
      "eval_accuracy": 0.8,
      "eval_loss": 0.866399347782135,
      "eval_runtime": 1.5559,
      "eval_samples_per_second": 38.564,
      "eval_steps_per_second": 1.285,
      "step": 263
    },
    {
      "epoch": 34.06,
      "grad_norm": 1.5001689195632935,
      "learning_rate": 5.681818181818182e-06,
      "loss": 0.0381,
      "step": 264
    },
    {
      "epoch": 34.97,
      "eval_accuracy": 0.8,
      "eval_loss": 0.8817654252052307,
      "eval_runtime": 1.5241,
      "eval_samples_per_second": 39.366,
      "eval_steps_per_second": 1.312,
      "step": 271
    },
    {
      "epoch": 35.61,
      "grad_norm": 0.628817617893219,
      "learning_rate": 3.409090909090909e-06,
      "loss": 0.0338,
      "step": 276
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9105542302131653,
      "eval_runtime": 1.5534,
      "eval_samples_per_second": 38.625,
      "eval_steps_per_second": 1.287,
      "step": 279
    },
    {
      "epoch": 36.9,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9056997299194336,
      "eval_runtime": 1.5384,
      "eval_samples_per_second": 39.002,
      "eval_steps_per_second": 1.3,
      "step": 286
    },
    {
      "epoch": 37.16,
      "grad_norm": 0.9884141683578491,
      "learning_rate": 1.1363636363636364e-06,
      "loss": 0.0443,
      "step": 288
    },
    {
      "epoch": 37.94,
      "eval_accuracy": 0.7833333333333333,
      "eval_loss": 0.9011732339859009,
      "eval_runtime": 1.6397,
      "eval_samples_per_second": 36.593,
      "eval_steps_per_second": 1.22,
      "step": 294
    },
    {
      "epoch": 37.94,
      "step": 294,
      "total_flos": 2.864620236542755e+18,
      "train_loss": 0.20059432347818296,
      "train_runtime": 1595.666,
      "train_samples_per_second": 25.637,
      "train_steps_per_second": 0.184
    }
  ],
  "logging_steps": 12,
  "max_steps": 294,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 42,
  "save_steps": 500,
  "total_flos": 2.864620236542755e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}