llama.cpp
2ceda3f6 - ggml-cpu: use LUT for converting e8->f32 scales on x86 (#19288)

Commit

4 days ago

ggml-cpu: use LUT for converting e8->f32 scales on x86 (#19288) * ggml-cpu: use LUT for converting e8->f32 scales on x86 * add dispatch based on macro

References

#19288 - ggml-cpu: use LUT for converting e8->f32 scales on x86

Author

am17an

Parents

44008ce8

llama.cpp 2ceda3f6 - ggml-cpu: use LUT for converting e8->f32 scales on x86 (#19288)

llama.cpp
2ceda3f6 - ggml-cpu: use LUT for converting e8->f32 scales on x86 (#19288)