Just a project that implements the "chat" example from SciSharp/LLamaSharp
Since 0.5.1 GGUF format models are required, not GGML.
e.g. TheBloke/Llama-2-7b-Chat-GGUF
Requires a GPU with enough RAM and CUDA support
-
Comment out the CPU backend package (
LLamaSharp.Backend.Cpu) in the.csproj