checkpoint-1564/trainer_state.json

{
  "best_global_step": 1564,
  "best_metric": 0.3568730354309082,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1564",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1564,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.639386189258312,
      "grad_norm": 4.334684371948242,
      "learning_rate": 0.000160801393728223,
      "loss": 0.642,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.74792,
      "eval_loss": 0.5220636129379272,
      "eval_model_preparation_time": 0.0024,
      "eval_runtime": 325.9979,
      "eval_samples_per_second": 76.688,
      "eval_steps_per_second": 9.586,
      "step": 782
    },
    {
      "epoch": 1.278772378516624,
      "grad_norm": 6.341824054718018,
      "learning_rate": 0.00011724738675958189,
      "loss": 0.554,
      "step": 1000
    },
    {
      "epoch": 1.918158567774936,
      "grad_norm": 3.435889482498169,
      "learning_rate": 7.369337979094078e-05,
      "loss": 0.4517,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8422,
      "eval_loss": 0.3568730354309082,
      "eval_model_preparation_time": 0.0024,
      "eval_runtime": 325.7227,
      "eval_samples_per_second": 76.752,
      "eval_steps_per_second": 9.594,
      "step": 1564
    }
  ],
  "logging_steps": 500,
  "max_steps": 2346,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 4.1877747910656e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}