checkpoint-1875/trainer_state.json

{
  "best_global_step": 1875,
  "best_metric": 0.027827957645058632,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1875",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1875,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.26666666666666666,
      "grad_norm": 10.099852561950684,
      "learning_rate": 7.534246575342466e-06,
      "loss": 0.3566,
      "step": 500
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 5.800719738006592,
      "learning_rate": 4.7945205479452054e-06,
      "loss": 0.0881,
      "step": 1000
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.3786800503730774,
      "learning_rate": 2.0547945205479454e-06,
      "loss": 0.0549,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9927833333333334,
      "eval_loss": 0.027827957645058632,
      "eval_model_preparation_time": 0.0023,
      "eval_runtime": 739.6755,
      "eval_samples_per_second": 81.117,
      "eval_steps_per_second": 10.14,
      "step": 1875
    }
  ],
  "logging_steps": 500,
  "max_steps": 1875,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "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": 5.02566312665088e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}