Show HN: 3.125-Bit LLM quantization bypassing tensor cores

3 points | by dmaniss 5 hours ago

1 comments