trainer_state.json

{
  "best_metric": 3.0357704162597656,
  "best_model_checkpoint": "clip-br-10k/checkpoint-626",
  "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": 3.076430082321167,
      "eval_runtime": 267.8742,
      "eval_samples_per_second": 74.662,
      "eval_steps_per_second": 2.333,
      "step": 313
    },
    {
      "epoch": 1.6,
      "learning_rate": 2.3375931842385517e-05,
      "loss": 2.5288,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 3.0357704162597656,
      "eval_runtime": 269.9693,
      "eval_samples_per_second": 74.083,
      "eval_steps_per_second": 2.315,
      "step": 626
    },
    {
      "epoch": 3.0,
      "eval_loss": 3.234591007232666,
      "eval_runtime": 269.0211,
      "eval_samples_per_second": 74.344,
      "eval_steps_per_second": 2.323,
      "step": 939
    },
    {
      "epoch": 3.0,
      "step": 939,
      "total_flos": 3012083735040000.0,
      "train_loss": 1.8548644759530584,
      "train_runtime": 1620.6041,
      "train_samples_per_second": 18.512,
      "train_steps_per_second": 0.579
    }
  ],
  "max_steps": 939,
  "num_train_epochs": 3,
  "total_flos": 3012083735040000.0,
  "trial_name": null,
  "trial_params": null
}