REVIEW 3 cited by
Kangaroo: Lossless Self-Speculative Decoding via Double Early Exiting
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Speculative decoding has demonstrated its effectiveness in accelerating the inference of large language models while maintaining a consistent sampling distribution. However, the conventional approach of training a separate draft model to achieve a satisfactory token acceptance rate can be costly. Drawing inspiration from early exiting, we propose a novel self-speculative decoding framework \emph{Kangaroo}, which uses a fixed shallow sub-network as a self-draft model, with the remaining layers serving as the larger target model. We train a lightweight and efficient adapter module on top of the sub-network to bridge the gap between the sub-network and the full model's representation ability. It is noteworthy that the inference latency of the self-draft model may no longer be negligible compared to the large model, necessitating strategies to increase the token acceptance rate while minimizing the drafting steps of the small model. To address this challenge, we introduce an additional early exiting mechanism for generating draft tokens. Specifically, we halt the small model's subsequent prediction during the drafting phase once the confidence level for the current token falls below a certain threshold. Extensive experiments on the Spec-Bench demonstrate the effectiveness of Kangaroo. Under single-sequence verification, Kangaroo achieves speedups up to $1.68\times$ on Spec-Bench, outperforming Medusa-1 with 88.7\% fewer additional parameters (67M compared to 591M). The code for Kangaroo is available at https://github.com/Equationliu/Kangaroo.
Forward citations
Cited by 3 Pith papers
-
Approximate Speculative Decoding
A training-free verifier, ASD, allows a bounded number of low-regret draft-target mismatches during speculative decoding and reuses the target-greedy suffix, yielding 3 to 15 percent throughput gains with mostly small...
-
AdaDecode: Accelerating LLM Decoding with Adaptive Layer Parallelism
AdaDecode speeds up LLM generation by predicting tokens at early layers when confidence is high, running the skipped layers in parallel, and verifying the output exactly matches standard decoding.
-
CLaSp: In-Context Layer Skip for Self-Speculative Decoding
A training-free, context-adaptive layer-skipping method for self-speculative decoding that reports roughly 1.1x to 1.8x speedups on LLaMA models while preserving output distribution.
Discussion (0). Continue with ORCID to comment.