checkpoint-4500/trainer_state.json

{
  "best_metric": 0.10008788853883743,
  "best_model_checkpoint": "clip_base_noaug/checkpoint-4500",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 4500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 34.92070007324219,
      "learning_rate": 8.988764044943819e-07,
      "loss": 0.2967,
      "step": 500
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 45.73786163330078,
      "learning_rate": 7.865168539325843e-07,
      "loss": 0.1592,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "grad_norm": 26.203510284423828,
      "learning_rate": 6.741573033707865e-07,
      "loss": 0.1429,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.94675,
      "eval_loss": 0.14425331354141235,
      "eval_model_preparation_time": 0.004,
      "eval_runtime": 87.8205,
      "eval_samples_per_second": 273.285,
      "eval_steps_per_second": 8.54,
      "step": 1500
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 42.80449676513672,
      "learning_rate": 5.617977528089887e-07,
      "loss": 0.123,
      "step": 2000
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 37.18833541870117,
      "learning_rate": 4.4943820224719097e-07,
      "loss": 0.1134,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "grad_norm": 57.78050231933594,
      "learning_rate": 3.3707865168539325e-07,
      "loss": 0.1023,
      "step": 3000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9605833333333333,
      "eval_loss": 0.10788565874099731,
      "eval_model_preparation_time": 0.004,
      "eval_runtime": 101.8398,
      "eval_samples_per_second": 235.664,
      "eval_steps_per_second": 7.365,
      "step": 3000
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 44.27554702758789,
      "learning_rate": 2.2471910112359549e-07,
      "loss": 0.0957,
      "step": 3500
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 29.09344482421875,
      "learning_rate": 1.1235955056179774e-07,
      "loss": 0.0897,
      "step": 4000
    },
    {
      "epoch": 3.0,
      "grad_norm": 18.151649475097656,
      "learning_rate": 0.0,
      "loss": 0.0888,
      "step": 4500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.962375,
      "eval_loss": 0.10008788853883743,
      "eval_model_preparation_time": 0.004,
      "eval_runtime": 88.2967,
      "eval_samples_per_second": 271.811,
      "eval_steps_per_second": 8.494,
      "step": 4500
    }
  ],
  "logging_steps": 500,
  "max_steps": 4500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.273880039240499e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}