llama.cpp
llama.cpp copied to clipboard
[SYCL] add conv support
- [x] I have read the contributing guidelines
- Self-reported review complexity:
- [x] Low
- [ ] Medium
- [ ] High
To support SYCL backend for stablediffusion.cpp https://github.com/leejet/stable-diffusion.cpp/issues/308 This is a temporary solution, we will introduce OneDNN for convolution performance