hf_quant_config.json

{
  "producer": {
    "name": "glm4v-nvfp4-layer-by-layer",
    "version": "1.0.0"
  },
  "quantization": {
    "quant_algo": "NVFP4",
    "kv_cache_quant_algo": null,
    "group_size": 16
  }
}