llama.cpp
5cca2542 - convert : avoid dequantizing mxfp4 for GPT-OSS (#16756)

Commit

59 days ago

convert : avoid dequantizing mxfp4 for GPT-OSS (#16756)

References

#16756 - convert : avoid dequantizing mxfp4 for GPT-OSS

Author

compilade

compilade

Parents

Loading