trainer_state.json

{
  "best_metric": 0.21961037814617157,
  "best_model_checkpoint": "./vit-base-highways-2/checkpoint-100",
  "epoch": 2.0,
  "global_step": 126,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16,
      "learning_rate": 0.00018412698412698412,
      "loss": 0.0048,
      "step": 10
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00016825396825396826,
      "loss": 0.0016,
      "step": 20
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00015238095238095237,
      "loss": 0.1014,
      "step": 30
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001365079365079365,
      "loss": 0.2578,
      "step": 40
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00012063492063492063,
      "loss": 0.4379,
      "step": 50
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00010476190476190477,
      "loss": 0.2386,
      "step": 60
    },
    {
      "epoch": 1.11,
      "learning_rate": 8.888888888888889e-05,
      "loss": 0.0062,
      "step": 70
    },
    {
      "epoch": 1.27,
      "learning_rate": 7.301587301587302e-05,
      "loss": 0.0102,
      "step": 80
    },
    {
      "epoch": 1.43,
      "learning_rate": 5.714285714285714e-05,
      "loss": 0.0596,
      "step": 90
    },
    {
      "epoch": 1.59,
      "learning_rate": 4.126984126984127e-05,
      "loss": 0.0012,
      "step": 100
    },
    {
      "epoch": 1.59,
      "eval_accuracy": 0.96,
      "eval_loss": 0.21961037814617157,
      "eval_runtime": 32.5474,
      "eval_samples_per_second": 3.841,
      "eval_steps_per_second": 0.492,
      "step": 100
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.5396825396825397e-05,
      "loss": 0.001,
      "step": 110
    },
    {
      "epoch": 1.9,
      "learning_rate": 9.523809523809523e-06,
      "loss": 0.0175,
      "step": 120
    },
    {
      "epoch": 2.0,
      "step": 126,
      "total_flos": 7.7491989614592e+16,
      "train_loss": 0.09034528125018354,
      "train_runtime": 299.6231,
      "train_samples_per_second": 3.338,
      "train_steps_per_second": 0.421
    }
  ],
  "max_steps": 126,
  "num_train_epochs": 2,
  "total_flos": 7.7491989614592e+16,
  "trial_name": null,
  "trial_params": null
}