checkpoint-1022/trainer_state.json

{
  "best_global_step": 1022,
  "best_metric": 0.08018568158149719,
  "best_model_checkpoint": "siglip2-finetune-full/checkpoint-1022",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1022,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9784735812133072,
      "grad_norm": 10.443733215332031,
      "learning_rate": 1.0740740740740742e-05,
      "loss": 0.3064,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9588865096359743,
      "eval_loss": 0.11907291412353516,
      "eval_model_preparation_time": 0.0024,
      "eval_runtime": 205.0429,
      "eval_samples_per_second": 79.715,
      "eval_steps_per_second": 9.969,
      "step": 511
    },
    {
      "epoch": 1.9569471624266144,
      "grad_norm": 3.556680917739868,
      "learning_rate": 4.526748971193416e-07,
      "loss": 0.1325,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9727745487916795,
      "eval_loss": 0.08018568158149719,
      "eval_model_preparation_time": 0.0024,
      "eval_runtime": 203.1892,
      "eval_samples_per_second": 80.442,
      "eval_steps_per_second": 10.06,
      "step": 1022
    }
  ],
  "logging_steps": 500,
  "max_steps": 1022,
  "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": 2.738057975951155e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}