Remove the stale nested text_config.quantization_config so inference engines detect Kimi-K2.6-NVFP4 from the top-level ModelOpt/NVFP4 metadata, matching Kimi-K2.5-NVFP4

NVIDIA org

Thanks for the feedback, we have config.json updated.

mmangkad changed pull request status to closed

Sign up or log in to comment