← 研踪 PaperTrailarXiv:2611.19896 · v1
LoopSkim: Lossless Multi-Depth Speculative Decoding for Looped Language Models
首次提交 2026-11-15 17:21
版本更新 2026-11-15 17:21
雷达观测 2026-11-17 12:06
摘要
Looped language models such as Ouro and Raven apply one weight-shared block several times per token, so autoregressive decoding (AR) runs every loop of every token in sequence. Pipelined self-speculative decoders overlap these loops by drafting tokens from shallow readouts and verifying them at full depth. Yet every loop of a looped LM ends in a readout of the next token, and draft tokens pass through every depth anyway, so each depth can check them without an extra pass. We introduce LoopSkim, in which the depth of the pipeline is the verifier. Draft tokens copied from the context, as in prompt lookup decoding, enter after the first loops, any later readout that contradicts one replaces it at that depth, and only full-depth readouts commit tokens. LoopSkim fixes the random draws of every position in advance, so each check compares two tokens and decoding stays lossless in exact arithmetic under sampling for any number of checks and draft chains. Without training, its speedup over AR averages 1.73x that of LoopSpec across greedy decoding and sampling, and its checks between the first loops and full depth speed it up on every model. An optional draft head adds a second chain and raises this average to 1.99x.
作者备注:23 pages, 3 figures, 16 tables
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-15 17:21 · 最新观测版本
LoopSkim: Lossless Multi-Depth Speculative Decoding for Looped Language Models
Looped language models such as Ouro and Raven apply one weight-shared block several times per token, so autoregressive decoding (AR) runs every loop of every token in sequence. Pipelined self-speculative decoders overlap these loops by drafting tokens from shallow readouts and verifying them at full depth. Yet every loop of a looped LM ends in a readout of the next token, and draft tokens pass through every depth anyway, so each depth can check them without an extra pass. We introduce LoopSkim, in which the depth of the pipeline is the verifier. Draft tokens copied from the context, as in prompt lookup decoding, enter after the first loops, any later readout that contradicts one replaces it at that depth, and only full-depth readouts commit tokens. LoopSkim fixes the random draws of every position in advance, so each check compares two tokens and decoding stays lossless in exact arithmetic under sampling for any number of checks and draft chains. Without training, its speedup over AR averages 1.73x that of LoopSpec across greedy decoding and sampling, and its checks between the first loops and full depth speed it up on every model. An optional draft head adds a second chain and raises this average to 1.99x.
查看此版本解读 →查看此版本原文 ↗