llama.cpp
a5114249 - [SYCL] Support Q4_1, Q5_0, Q5_1 in Flash-attention (#23812)

Commit

1 day ago

[SYCL] Support Q4_1, Q5_0, Q5_1 in Flash-attention (#23812) * support Q4_1, Q5_0, Q5_1 * update ut case

References

Author

arthw

Parents