Note that 2bit quantization has worse performance compared to 3bit quantization as shown in our paper. The inclusion of 2-bit quantization is just an extreme exploration about deploy LLM in mobile ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results