{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 19.997442455242968,
  "global_step": 3900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9097127318382263,
      "eval_loss": 0.2761473059654236,
      "eval_runtime": 4.9719,
      "eval_samples_per_second": 147.026,
      "eval_steps_per_second": 18.504,
      "step": 195
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9220246076583862,
      "eval_loss": 0.310268372297287,
      "eval_runtime": 4.9488,
      "eval_samples_per_second": 147.712,
      "eval_steps_per_second": 18.59,
      "step": 390
    },
    {
      "epoch": 2.56,
      "learning_rate": 4.087454207611332e-05,
      "loss": 0.2255,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9124487042427063,
      "eval_loss": 0.29651716351509094,
      "eval_runtime": 4.9678,
      "eval_samples_per_second": 147.148,
      "eval_steps_per_second": 18.519,
      "step": 585
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9233925938606262,
      "eval_loss": 0.3387739956378937,
      "eval_runtime": 4.9332,
      "eval_samples_per_second": 148.181,
      "eval_steps_per_second": 18.649,
      "step": 780
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9206566214561462,
      "eval_loss": 0.42386311292648315,
      "eval_runtime": 5.0121,
      "eval_samples_per_second": 145.847,
      "eval_steps_per_second": 18.356,
      "step": 975
    },
    {
      "epoch": 5.13,
      "learning_rate": 3.486358000609665e-05,
      "loss": 0.1137,
      "step": 1000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.43168818950653076,
      "eval_runtime": 4.956,
      "eval_samples_per_second": 147.497,
      "eval_steps_per_second": 18.563,
      "step": 1170
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9206566214561462,
      "eval_loss": 0.5392778515815735,
      "eval_runtime": 5.1536,
      "eval_samples_per_second": 141.844,
      "eval_steps_per_second": 17.852,
      "step": 1365
    },
    {
      "epoch": 7.69,
      "learning_rate": 2.885261793607999e-05,
      "loss": 0.0469,
      "step": 1500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9151846766471863,
      "eval_loss": 0.46178123354911804,
      "eval_runtime": 4.9372,
      "eval_samples_per_second": 148.059,
      "eval_steps_per_second": 18.634,
      "step": 1560
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9138166904449463,
      "eval_loss": 0.6065310835838318,
      "eval_runtime": 4.9633,
      "eval_samples_per_second": 147.282,
      "eval_steps_per_second": 18.536,
      "step": 1755
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9165526628494263,
      "eval_loss": 0.5557906031608582,
      "eval_runtime": 5.0575,
      "eval_samples_per_second": 144.538,
      "eval_steps_per_second": 18.191,
      "step": 1950
    },
    {
      "epoch": 10.26,
      "learning_rate": 2.2841655866063323e-05,
      "loss": 0.0283,
      "step": 2000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9247605800628662,
      "eval_loss": 0.5483647584915161,
      "eval_runtime": 4.8959,
      "eval_samples_per_second": 149.308,
      "eval_steps_per_second": 18.791,
      "step": 2145
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9151846766471863,
      "eval_loss": 0.7020018100738525,
      "eval_runtime": 5.0449,
      "eval_samples_per_second": 144.899,
      "eval_steps_per_second": 18.236,
      "step": 2340
    },
    {
      "epoch": 12.82,
      "learning_rate": 1.6830693796046662e-05,
      "loss": 0.015,
      "step": 2500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.7031099796295166,
      "eval_runtime": 4.8879,
      "eval_samples_per_second": 149.552,
      "eval_steps_per_second": 18.822,
      "step": 2535
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9151846766471863,
      "eval_loss": 0.7629717588424683,
      "eval_runtime": 5.0202,
      "eval_samples_per_second": 145.613,
      "eval_steps_per_second": 18.326,
      "step": 2730
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9165526628494263,
      "eval_loss": 0.7813730835914612,
      "eval_runtime": 4.9102,
      "eval_samples_per_second": 148.874,
      "eval_steps_per_second": 18.737,
      "step": 2925
    },
    {
      "epoch": 15.38,
      "learning_rate": 1.0819731726029997e-05,
      "loss": 0.0043,
      "step": 3000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.7715274691581726,
      "eval_runtime": 5.0959,
      "eval_samples_per_second": 143.449,
      "eval_steps_per_second": 18.054,
      "step": 3120
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.7615724802017212,
      "eval_runtime": 4.852,
      "eval_samples_per_second": 150.659,
      "eval_steps_per_second": 18.961,
      "step": 3315
    },
    {
      "epoch": 17.95,
      "learning_rate": 4.808769656013331e-06,
      "loss": 0.0031,
      "step": 3500
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9220246076583862,
      "eval_loss": 0.7719457745552063,
      "eval_runtime": 5.0863,
      "eval_samples_per_second": 143.72,
      "eval_steps_per_second": 18.088,
      "step": 3510
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.779067873954773,
      "eval_runtime": 4.9894,
      "eval_samples_per_second": 146.509,
      "eval_steps_per_second": 18.439,
      "step": 3705
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.7817478775978088,
      "eval_runtime": 4.8884,
      "eval_samples_per_second": 149.538,
      "eval_steps_per_second": 18.82,
      "step": 3900
    },
    {
      "epoch": 20.0,
      "step": 3900,
      "total_flos": 4114604473061376.0,
      "train_loss": 0.056036626369907305,
      "train_runtime": 3116.2149,
      "train_samples_per_second": 20.076,
      "train_steps_per_second": 1.252
    }
  ],
  "max_steps": 3900,
  "num_train_epochs": 20,
  "total_flos": 4114604473061376.0,
  "trial_name": null,
  "trial_params": null
}