1 paper
Kecheng Chen, Yuyang He, Cheng Gong +8
Block diffusion accelerates speculative decoding by drafting multiple tokens in one forward pass. However, each position predicts a marginal distribution without observing earlier…