hf_quant_config.json

{
  "producer": {
    "name": "glm53-dflash2-mxfp8-converter",
    "version": 1
  },
  "quantization": {
    "exclude_modules": [
      "lm_head"
    ],
    "kv_cache_quant_algo": null,
    "quant_algo": "MXFP8"
  }
}