Abstract: Vector quantization (VQ), which treats a vector as a compression unit, gains increasing research interests for its potential to accelerate large language models (LLMs). Compared to ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback