checkpoint-1128/trainer_state.json

{
  "best_global_step": 1128,
  "best_metric": 0.3800027370452881,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1128",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1128,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8865248226950354,
      "grad_norm": 4.772475242614746,
      "learning_rate": 0.00011651205936920225,
      "loss": 0.8994,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7787571373135983,
      "eval_loss": 0.5819662809371948,
      "eval_model_preparation_time": 0.0023,
      "eval_runtime": 243.228,
      "eval_samples_per_second": 74.165,
      "eval_steps_per_second": 9.271,
      "step": 564
    },
    {
      "epoch": 1.773049645390071,
      "grad_norm": 7.929209232330322,
      "learning_rate": 2.3747680890538035e-05,
      "loss": 0.6065,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8589167913964189,
      "eval_loss": 0.3800027370452881,
      "eval_model_preparation_time": 0.0023,
      "eval_runtime": 243.5063,
      "eval_samples_per_second": 74.08,
      "eval_steps_per_second": 9.261,
      "step": 1128
    }
  ],
  "logging_steps": 500,
  "max_steps": 1128,
  "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": 3.0218059506714993e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}