llama.cpp
5f4001e0 - rm cuda supports_op guard, fix webgpu clang-format

Commit

1 day ago

rm cuda supports_op guard, fix webgpu clang-format

References

xsn/ggml_rope_off

#27120 - ggml: add ggml_rope_set_offset (+ metal support)

Author

ngxson

ngxson

Parents

Loading