RedHatAI

/

diffusiongemma-26B-A4B-it-FP8-dynamic

like 11

Follow

Red Hat AI 2.85k

Hi Red Hat AI team,

Thanks for releasing the DiffusionGemma FP8/NVFP4 checkpoints and the quantization scripts!

I noticed the examples import:

from llmcompressor.modeling.diffusion_gemma4 import CalibrationDiffusionGemmaTextExperts

but this module doesn't seem to be available in the public llm-compressor repository or any released version. Would it be possible to share the implementation (or the equivalent patch)? It would greatly help the community reproduce the quantization workflow for DiffusionGemma.

Thanks!