When Parallel Drafter Meets Parallel Speculative Decoding
Read the original on arXiv Computation and Language →The paper introduces DPara, a parallel speculative decoding framework that builds on DSpark-style parallel drafters. DPara eliminates the need for probabilistic guesses by precomputing draft representations for every acceptance boundary and using a lightweight autoregressive head to combine verification outcomes with these representations, enabling full parallelization of the backbone forward pass. Experiments on Qwen3-8B and Qwen3-14B across multiple benchmarks demonstrate average speedups of 3.21× and 3.52× over autoregressive decoding, outperforming existing serial and parallel speculative decoding methods.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.