3 ms·
34B Q4 will use around 20GB of memory. If it's running slow, make sure metal is actually being used[0]. You can get as much as a 50-100% boost in tokens/s, if
by SparkyMcUnicorn 3y ago
34B Q4 will use around 20GB of memory.
If it's running slow, make sure metal is actually being used[0]. You can get as much as a 50-100% boost in tokens/s, if by chance it's not enabled.
I'm averaging 7 to 8 tokens/s on an M1 Max 10 core (24 GPU cores).
[0] if using llama-cpp-python (or text-generation-webui, ollama, etc) try:
`pip uninstall llama-cpp-python && CMAKE_ARGS="-DLLAMA_METAL=on" FORCE_CMAKE=1 pip install llama-cpp-python`
- npsomaratna 3y agoThank you. I had to reduce the context length to get this to work without crashing (from 16k to 8k)—and I'm seeing the ~100% speed up you mentioned. However, when I run the LLM, OSX becomes sluggish. I assume this is because the GPU's utilized to the point where hardware-based rendering slows down due to insufficient resources. I wonder if there's a way to avoid that slowdown?
- SparkyMcUnicorn 3y agoI haven't noticed any slowdowns. Maybe check that threads/n_threads is set correctly for your machine (total cores - 2. 10 cores = 8, 8 cores = 6). n_gpu_layers should also be set to anything other than 0 (default). I don't think the exact number matters for metal, but I use 128.