checkpoint-928/trainer_state.json

{
  "best_global_step": 928,
  "best_metric": 1.0770337581634521,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-928",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 928,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6359832635983264,
      "eval_loss": 2.0701308250427246,
      "eval_model_preparation_time": 0.0031,
      "eval_runtime": 215.6864,
      "eval_samples_per_second": 68.702,
      "eval_steps_per_second": 8.591,
      "step": 464
    },
    {
      "epoch": 1.0775862068965518,
      "grad_norm": 49.15087890625,
      "learning_rate": 2.4052863436123347e-06,
      "loss": 3.5107,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8090160615467674,
      "eval_loss": 1.0770337581634521,
      "eval_model_preparation_time": 0.0031,
      "eval_runtime": 213.0035,
      "eval_samples_per_second": 69.567,
      "eval_steps_per_second": 8.699,
      "step": 928
    }
  ],
  "logging_steps": 500,
  "max_steps": 2320,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 2.4847302002090803e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}