Bonsai 8B: a 1-bit LLM that fits in 1.15GB(firethering.com)
firethering.com
Bonsai 8B: a 1-bit LLM that fits in 1.15GB
https://firethering.com/bonsai-8b-1bit-llm/
2 comments
If this holds, does it unlock 100B+ models running locally in ~tens of GB RAM? Or does accuracy collapse before that point?