{
  "best_metric": 0.49532002210617065,
  "best_model_checkpoint": "./vit-dropout-0.4/checkpoint-2889",
  "epoch": 19.0,
  "eval_steps": 500,
  "global_step": 6099,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 20.27271842956543,
      "learning_rate": 9.008498583569406e-05,
      "loss": 1.5638,
      "step": 321
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.625866851595007,
      "eval_f1": 0.6634599560471133,
      "eval_loss": 0.917059063911438,
      "eval_precision": 0.7843432532062684,
      "eval_recall": 0.625866851595007,
      "eval_runtime": 22.9894,
      "eval_samples_per_second": 125.449,
      "eval_steps_per_second": 15.703,
      "step": 321
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.8246476650238037,
      "learning_rate": 9.967994652315143e-05,
      "loss": 1.1092,
      "step": 642
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7007628294036061,
      "eval_f1": 0.7192900679194619,
      "eval_loss": 0.7738657593727112,
      "eval_precision": 0.7903482253631083,
      "eval_recall": 0.7007628294036061,
      "eval_runtime": 22.8225,
      "eval_samples_per_second": 126.367,
      "eval_steps_per_second": 15.818,
      "step": 642
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.247750282287598,
      "learning_rate": 9.856371224544644e-05,
      "loss": 0.9892,
      "step": 963
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7780859916782247,
      "eval_f1": 0.778423467149636,
      "eval_loss": 0.6145934462547302,
      "eval_precision": 0.790423999944827,
      "eval_recall": 0.7780859916782247,
      "eval_runtime": 22.6731,
      "eval_samples_per_second": 127.199,
      "eval_steps_per_second": 15.922,
      "step": 963
    },
    {
      "epoch": 4.0,
      "grad_norm": 13.25033950805664,
      "learning_rate": 9.666449571439611e-05,
      "loss": 0.8735,
      "step": 1284
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7454923717059639,
      "eval_f1": 0.7615557397273891,
      "eval_loss": 0.6262115836143494,
      "eval_precision": 0.807493921529863,
      "eval_recall": 0.7454923717059639,
      "eval_runtime": 22.9609,
      "eval_samples_per_second": 125.605,
      "eval_steps_per_second": 15.722,
      "step": 1284
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.960892677307129,
      "learning_rate": 9.401291758831328e-05,
      "loss": 0.8118,
      "step": 1605
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7163661581137309,
      "eval_f1": 0.7311152376922296,
      "eval_loss": 0.7255703806877136,
      "eval_precision": 0.8185184282178283,
      "eval_recall": 0.7163661581137309,
      "eval_runtime": 22.9078,
      "eval_samples_per_second": 125.896,
      "eval_steps_per_second": 15.759,
      "step": 1605
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.861917972564697,
      "learning_rate": 9.065172868916128e-05,
      "loss": 0.7794,
      "step": 1926
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7819001386962552,
      "eval_f1": 0.7924960521456311,
      "eval_loss": 0.608811616897583,
      "eval_precision": 0.8200552259690934,
      "eval_recall": 0.7819001386962552,
      "eval_runtime": 22.5121,
      "eval_samples_per_second": 128.109,
      "eval_steps_per_second": 16.036,
      "step": 1926
    },
    {
      "epoch": 7.0,
      "grad_norm": 8.067416191101074,
      "learning_rate": 8.663512074027613e-05,
      "loss": 0.6835,
      "step": 2247
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7624826629680999,
      "eval_f1": 0.778275921221607,
      "eval_loss": 0.5834646821022034,
      "eval_precision": 0.8169986016936643,
      "eval_recall": 0.7624826629680999,
      "eval_runtime": 23.3953,
      "eval_samples_per_second": 123.273,
      "eval_steps_per_second": 15.43,
      "step": 2247
    },
    {
      "epoch": 8.0,
      "grad_norm": 3.7579257488250732,
      "learning_rate": 8.202785264490498e-05,
      "loss": 0.5997,
      "step": 2568
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.765256588072122,
      "eval_f1": 0.7821382649754356,
      "eval_loss": 0.6475569605827332,
      "eval_precision": 0.826387274966978,
      "eval_recall": 0.765256588072122,
      "eval_runtime": 22.6325,
      "eval_samples_per_second": 127.427,
      "eval_steps_per_second": 15.95,
      "step": 2568
    },
    {
      "epoch": 9.0,
      "grad_norm": 10.09793758392334,
      "learning_rate": 7.692088630529957e-05,
      "loss": 0.5814,
      "step": 2889
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8401525658807212,
      "eval_f1": 0.8403607055183363,
      "eval_loss": 0.49532002210617065,
      "eval_precision": 0.8423956140728975,
      "eval_recall": 0.8401525658807212,
      "eval_runtime": 25.1073,
      "eval_samples_per_second": 114.867,
      "eval_steps_per_second": 14.378,
      "step": 2889
    },
    {
      "epoch": 10.0,
      "grad_norm": 7.585777282714844,
      "learning_rate": 7.136468478571602e-05,
      "loss": 0.5134,
      "step": 3210
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.8103328710124826,
      "eval_f1": 0.8220041299869107,
      "eval_loss": 0.5335156321525574,
      "eval_precision": 0.8534486868065884,
      "eval_recall": 0.8103328710124826,
      "eval_runtime": 24.7217,
      "eval_samples_per_second": 116.659,
      "eval_steps_per_second": 14.603,
      "step": 3210
    },
    {
      "epoch": 11.0,
      "grad_norm": 20.268178939819336,
      "learning_rate": 6.546402506345253e-05,
      "loss": 0.5109,
      "step": 3531
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.8124133148404993,
      "eval_f1": 0.8191514417244506,
      "eval_loss": 0.5497095584869385,
      "eval_precision": 0.8414741317511313,
      "eval_recall": 0.8124133148404993,
      "eval_runtime": 23.8572,
      "eval_samples_per_second": 120.886,
      "eval_steps_per_second": 15.132,
      "step": 3531
    },
    {
      "epoch": 12.0,
      "grad_norm": 3.8051164150238037,
      "learning_rate": 5.931404220195339e-05,
      "loss": 0.4073,
      "step": 3852
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8311373092926491,
      "eval_f1": 0.831620079776665,
      "eval_loss": 0.5754452347755432,
      "eval_precision": 0.8347554735526493,
      "eval_recall": 0.8311373092926491,
      "eval_runtime": 24.6815,
      "eval_samples_per_second": 116.849,
      "eval_steps_per_second": 14.626,
      "step": 3852
    },
    {
      "epoch": 13.0,
      "grad_norm": 1.562414526939392,
      "learning_rate": 5.30138910478373e-05,
      "loss": 0.3255,
      "step": 4173
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.8574895977808599,
      "eval_f1": 0.8525988995704912,
      "eval_loss": 0.5593818426132202,
      "eval_precision": 0.8620968298732496,
      "eval_recall": 0.8574895977808599,
      "eval_runtime": 24.0309,
      "eval_samples_per_second": 120.012,
      "eval_steps_per_second": 15.022,
      "step": 4173
    },
    {
      "epoch": 14.0,
      "grad_norm": 0.7332964539527893,
      "learning_rate": 4.666514757870119e-05,
      "loss": 0.3288,
      "step": 4494
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8332177531206657,
      "eval_f1": 0.8401776032043191,
      "eval_loss": 0.6329751014709473,
      "eval_precision": 0.8607345037298127,
      "eval_recall": 0.8332177531206657,
      "eval_runtime": 23.7481,
      "eval_samples_per_second": 121.441,
      "eval_steps_per_second": 15.201,
      "step": 4494
    },
    {
      "epoch": 15.0,
      "grad_norm": 5.967740535736084,
      "learning_rate": 4.038959286483306e-05,
      "loss": 0.2434,
      "step": 4815
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8606102635228848,
      "eval_f1": 0.861866915129384,
      "eval_loss": 0.5198509693145752,
      "eval_precision": 0.8645868176470374,
      "eval_recall": 0.8606102635228848,
      "eval_runtime": 23.048,
      "eval_samples_per_second": 125.13,
      "eval_steps_per_second": 15.663,
      "step": 4815
    },
    {
      "epoch": 16.0,
      "grad_norm": 19.630001068115234,
      "learning_rate": 3.424923705075116e-05,
      "loss": 0.2185,
      "step": 5136
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.858876560332871,
      "eval_f1": 0.860451647543394,
      "eval_loss": 0.5324673652648926,
      "eval_precision": 0.8647326759799664,
      "eval_recall": 0.858876560332871,
      "eval_runtime": 22.8629,
      "eval_samples_per_second": 126.143,
      "eval_steps_per_second": 15.79,
      "step": 5136
    },
    {
      "epoch": 17.0,
      "grad_norm": 0.011605961248278618,
      "learning_rate": 2.8362827388964596e-05,
      "loss": 0.1707,
      "step": 5457
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.8640776699029126,
      "eval_f1": 0.8597682191227608,
      "eval_loss": 0.5524240136146545,
      "eval_precision": 0.8638803900922614,
      "eval_recall": 0.8640776699029126,
      "eval_runtime": 22.516,
      "eval_samples_per_second": 128.087,
      "eval_steps_per_second": 16.033,
      "step": 5457
    },
    {
      "epoch": 18.0,
      "grad_norm": 25.897785186767578,
      "learning_rate": 2.2825269192267617e-05,
      "loss": 0.1702,
      "step": 5778
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8522884882108183,
      "eval_f1": 0.8551640876798166,
      "eval_loss": 0.5471655130386353,
      "eval_precision": 0.8612118088331184,
      "eval_recall": 0.8522884882108183,
      "eval_runtime": 22.9729,
      "eval_samples_per_second": 125.539,
      "eval_steps_per_second": 15.714,
      "step": 5778
    },
    {
      "epoch": 19.0,
      "grad_norm": 0.07399608939886093,
      "learning_rate": 1.7725843316558628e-05,
      "loss": 0.128,
      "step": 6099
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8751733703190014,
      "eval_f1": 0.8745515169095409,
      "eval_loss": 0.559937596321106,
      "eval_precision": 0.8758481332673859,
      "eval_recall": 0.8751733703190014,
      "eval_runtime": 22.9935,
      "eval_samples_per_second": 125.427,
      "eval_steps_per_second": 15.7,
      "step": 6099
    },
    {
      "epoch": 19.0,
      "step": 6099,
      "total_flos": 7.550537882380222e+18,
      "train_loss": 0.579371564758471,
      "train_runtime": 2226.9019,
      "train_samples_per_second": 230.275,
      "train_steps_per_second": 14.415
    }
  ],
  "logging_steps": 500,
  "max_steps": 32100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 500,
  "total_flos": 7.550537882380222e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}