checkpoint-3000/trainer_state.json

{
  "best_metric": 0.11322234570980072,
  "best_model_checkpoint": "ai_vs_human_generated_image_detection/checkpoint-3000",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 0.953800618648529,
      "learning_rate": 2.5423728813559323e-06,
      "loss": 0.4979,
      "step": 500
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 1.084754228591919,
      "learning_rate": 2.033898305084746e-06,
      "loss": 0.146,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "grad_norm": 12.670453071594238,
      "learning_rate": 1.5254237288135592e-06,
      "loss": 0.1073,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9691056910569106,
      "eval_loss": 0.14485515654087067,
      "eval_model_preparation_time": 0.0038,
      "eval_runtime": 428.2517,
      "eval_samples_per_second": 74.676,
      "eval_steps_per_second": 9.336,
      "step": 1500
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 2.258631467819214,
      "learning_rate": 1.016949152542373e-06,
      "loss": 0.0859,
      "step": 2000
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 4.437932014465332,
      "learning_rate": 5.084745762711865e-07,
      "loss": 0.0785,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.19989100098609924,
      "learning_rate": 0.0,
      "loss": 0.0762,
      "step": 3000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9722639149468417,
      "eval_loss": 0.11322234570980072,
      "eval_model_preparation_time": 0.0038,
      "eval_runtime": 432.331,
      "eval_samples_per_second": 73.971,
      "eval_steps_per_second": 9.248,
      "step": 3000
    }
  ],
  "logging_steps": 500,
  "max_steps": 3000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 7.434581483623956e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}