3 ms·
It inherits from ARM NEON optimizations in llama.cpp but I believe it should also work on x86.
by antimatter15 4y ago
It inherits from ARM NEON optimizations in llama.cpp but I believe it should also work on x86.
- gpderetta 4y agoit built in a couple of seconds with no issues on my x86 machine. Downloading weights now.
- gorbypark 4y agoSomeone contributed AVX2 support a few days after the initial release, so llama.cpp (and the alpaca.cpp fork) should work on most x86 CPUs as well now.
- inaka10100101 4y agoCan we run it in colab. It isn't working for me.