AgntroAI/tq2-quaternary — GitHub trending stats & insights

Reach 125K+ monthly visitors

AgntroAI/tq2-quaternary

2-bit MoE expert quantization that beats same-size community quants on both speed and quality — by reclaiming TQ2_0's unused 4th level (ternary → quaternary).

Visit GitHub

Python

1 contributors

Custom license

website

Social mentions

Recent discussions about this repository across the web

⚠️ One hazard if you use these: never requantize or LoRA-merge into a 4-level GGUF. llama.cpp's from_float path clamps to [-1,1] and silently strips the 4th level. Run them as-is; don't feed them…

@AgntroAI · x.com

Repository activities

repository's daily and monthly activities across stars, forks, merged PRs, issues, and closed issues

Data is not available yet

Recent activity data for stars, forks, merged PRs, issues, and closed issues will appear here once available