{
  "best_global_step": 3227,
  "best_metric": 1.9226315021514893,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-3227",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 3227,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1549426712116517,
      "grad_norm": 46.12940979003906,
      "learning_rate": 2.7891942535915053e-06,
      "loss": 3.3109,
      "step": 500
    },
    {
      "epoch": 0.3098853424233034,
      "grad_norm": 61.79978561401367,
      "learning_rate": 2.5549656464709556e-06,
      "loss": 2.5591,
      "step": 1000
    },
    {
      "epoch": 0.4648280136349551,
      "grad_norm": 53.559627532958984,
      "learning_rate": 2.320737039350406e-06,
      "loss": 2.349,
      "step": 1500
    },
    {
      "epoch": 0.6197706848466068,
      "grad_norm": 60.83781433105469,
      "learning_rate": 2.0865084322298563e-06,
      "loss": 2.2111,
      "step": 2000
    },
    {
      "epoch": 0.7747133560582584,
      "grad_norm": 53.31537628173828,
      "learning_rate": 1.8522798251093069e-06,
      "loss": 2.0793,
      "step": 2500
    },
    {
      "epoch": 0.9296560272699101,
      "grad_norm": 51.50001907348633,
      "learning_rate": 1.6180512179887572e-06,
      "loss": 2.0166,
      "step": 3000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.45531234866828085,
      "eval_loss": 1.9226315021514893,
      "eval_model_preparation_time": 0.0024,
      "eval_runtime": 1344.5655,
      "eval_samples_per_second": 76.791,
      "eval_steps_per_second": 9.599,
      "step": 3227
    }
  ],
  "logging_steps": 500,
  "max_steps": 6454,
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 8.657435909896704e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}