checkpoint-4500/trainer_state.json

{
  "best_metric": 0.6481580138206482,
  "best_model_checkpoint": "vit_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": 1.2217068672180176,
      "learning_rate": 8.988764044943819e-07,
      "loss": 0.6327,
      "step": 500
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 3.5019662380218506,
      "learning_rate": 7.865168539325843e-07,
      "loss": 0.3976,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "grad_norm": 2.137709379196167,
      "learning_rate": 6.741573033707865e-07,
      "loss": 0.2865,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6211666666666666,
      "eval_loss": 0.6675366759300232,
      "eval_model_preparation_time": 0.0048,
      "eval_runtime": 212.8762,
      "eval_samples_per_second": 112.742,
      "eval_steps_per_second": 3.523,
      "step": 1500
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 1.9651539325714111,
      "learning_rate": 5.617977528089887e-07,
      "loss": 0.2452,
      "step": 2000
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 4.475180625915527,
      "learning_rate": 4.4943820224719097e-07,
      "loss": 0.2276,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.077700614929199,
      "learning_rate": 3.3707865168539325e-07,
      "loss": 0.2112,
      "step": 3000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.641625,
      "eval_loss": 0.6539896130561829,
      "eval_model_preparation_time": 0.0048,
      "eval_runtime": 195.4432,
      "eval_samples_per_second": 122.798,
      "eval_steps_per_second": 3.837,
      "step": 3000
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 4.195696830749512,
      "learning_rate": 2.2471910112359549e-07,
      "loss": 0.2018,
      "step": 3500
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 4.716664791107178,
      "learning_rate": 1.1235955056179774e-07,
      "loss": 0.1973,
      "step": 4000
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.225037574768066,
      "learning_rate": 0.0,
      "loss": 0.1967,
      "step": 4500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.648375,
      "eval_loss": 0.6481580138206482,
      "eval_model_preparation_time": 0.0048,
      "eval_runtime": 195.3219,
      "eval_samples_per_second": 122.874,
      "eval_steps_per_second": 3.84,
      "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": 2.2317693009002496e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}