Hello,
Is it possible to add the full 16 quantized version?
Some quantized model performance results are not good enough. We just added the native mxfp4 version. Enjoy!
Jinx Team,Best
Β· Sign up or log in to comment