llama.cpp
622cd010 - ggml: CUDA: add head size 72 for flash-attn (#16962)

Commit

2 days ago

ggml: CUDA: add head size 72 for flash-attn (#16962)

References

#16962 - CUDA: add head size 72 for flash-attn

Author

theo77186

theo77186

Parents

Loading