trainer_state.json

{
  "best_metric": 3.327958106994629,
  "best_model_checkpoint": "clip-roberta-10k/checkpoint-939",
  "epoch": 3.0,
  "global_step": 939,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 3.4277737140655518,
      "eval_runtime": 117.2709,
      "eval_samples_per_second": 85.273,
      "eval_steps_per_second": 2.669,
      "step": 313
    },
    {
      "epoch": 1.6,
      "learning_rate": 2.3375931842385517e-05,
      "loss": 3.4415,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 3.3748862743377686,
      "eval_runtime": 131.3188,
      "eval_samples_per_second": 76.151,
      "eval_steps_per_second": 2.384,
      "step": 626
    },
    {
      "epoch": 3.0,
      "eval_loss": 3.327958106994629,
      "eval_runtime": 131.0009,
      "eval_samples_per_second": 76.335,
      "eval_steps_per_second": 2.389,
      "step": 939
    },
    {
      "epoch": 3.0,
      "step": 939,
      "total_flos": 4005518400000000.0,
      "train_loss": 3.383876552723975,
      "train_runtime": 1383.4982,
      "train_samples_per_second": 21.684,
      "train_steps_per_second": 0.679
    }
  ],
  "max_steps": 939,
  "num_train_epochs": 3,
  "total_flos": 4005518400000000.0,
  "trial_name": null,
  "trial_params": null
}