checkpoint-1012/trainer_state.json

{
  "best_global_step": 1012,
  "best_metric": 0.18154847621917725,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1012",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1012,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9881422924901185,
      "grad_norm": 7.464842796325684,
      "learning_rate": 0.0002661122661122661,
      "loss": 1.5544,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6886938245657415,
      "eval_loss": 0.9519435167312622,
      "eval_model_preparation_time": 0.0026,
      "eval_runtime": 216.1541,
      "eval_samples_per_second": 74.84,
      "eval_steps_per_second": 9.359,
      "step": 506
    },
    {
      "epoch": 1.9762845849802373,
      "grad_norm": 4.09387731552124,
      "learning_rate": 6.237006237006238e-06,
      "loss": 0.3209,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.942510972368177,
      "eval_loss": 0.18154847621917725,
      "eval_model_preparation_time": 0.0026,
      "eval_runtime": 211.4297,
      "eval_samples_per_second": 76.512,
      "eval_steps_per_second": 9.568,
      "step": 1012
    }
  ],
  "logging_steps": 500,
  "max_steps": 1012,
  "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": 2.7104320294495396e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}