← Back to context Comment by 7speter 16 hours ago You can offload the vision projector to CPU/sysRAM 2 comments 7speter Reply slim 12 hours ago Awesome. How would you setup that ? skirmish 11 hours ago --no-mmproj-offload See the documentation: https://github.com/ggml-org/llama.cpp/blob/master/docs/multi...
slim 12 hours ago Awesome. How would you setup that ? skirmish 11 hours ago --no-mmproj-offload See the documentation: https://github.com/ggml-org/llama.cpp/blob/master/docs/multi...
skirmish 11 hours ago --no-mmproj-offload See the documentation: https://github.com/ggml-org/llama.cpp/blob/master/docs/multi...
Awesome. How would you setup that ?
See the documentation: https://github.com/ggml-org/llama.cpp/blob/master/docs/multi...