quantization_info.json

{
  "model_id": "/workspace/LlamaFactory_new/saves/qwen35-9b/merged/31d",
  "output_dir": "/workspace/LlamaFactory_new/saves/qwen35-9b/int8/31d",
  "tool": "llm-compressor",
  "method": "int8",
  "scheme": "W8A8",
  "calib_data": "data/splits/train_260421_total.json",
  "calib_samples": 4096,
  "max_seq_length": 512,
  "elapsed_load_s": 23.5,
  "elapsed_quant_s": 2.9,
  "elapsed_save_s": 16.6,
  "total_elapsed_s": 44.6
}