https://github.com/pitbox46/NanoQuant TLDR: NanoQuant is a quantization method to create 2 bit/weight, 1 bit/weight, 0.5 bit/weight, etc, quants of dense transformer models. I've followed the paper's methods and created my own implementation which is still very much a work in progress, but currently seems very promising. I am not affiliated with the NanoQuant team What is NanoQuant NanoQuant (Chong et al, 2026, https://arxiv.org/abs/2602.06694 ) is a post-training quantization method w...