minipasila

joined 1 year ago
[–] [email protected] 2 points 1 year ago

I don't know about that, but you could try GGML (llama.cpp). It has quantization up to 2-bits so that might be small enough.