Ill_Initiative_8793

joined 1 year ago
[–] Ill_Initiative_8793@alien.top 1 points 11 months ago

llama.cpp and upload some layers to VRAM, you may be able to run 70B, depends on quantization.