A gated LoRA and a small sampler let an autoregressive LLM draft multiple future tokens per step, and self-speculative verification converts those drafts into up to roughly 5x fewer generation steps.
SDSAT: Accelerating LLM Inference through Speculative Decoding with Semantic Adaptive Tokens
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We propose an acceleration scheme for large language models (LLMs) through Speculative Decoding with Semantic Adaptive Tokens (SDSAT). The primary objective of this design is to enhance the LLM model's ability to generate draft tokens more accurately without compromising the model's accuracy. The core strategies involve: 1) Fine-tune the model by incorporating semantic adaptive tokens that possess flexible decoding capabilities without changing its structure, allowing them to generate high-quality draft tokens. 2) By employing a training method that does not affect the standard tokens, the model can acquire parallel decoding abilities atop its original framework with minimal training overhead. 3) We have designed the "two-step-draft-then-verify" generation strategies using both greedy search and nucleus sampling. Experiments conducted on the CodeLlama-13B and 7B models have yielded speed increases of over 3.5X and 3.0X, respectively. Please refer to https://github.com/hasuoshenyun/SDSAT.
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Your LLM Knows the Future: Uncovering Its Multi-Token Prediction Potential
A gated LoRA and a small sampler let an autoregressive LLM draft multiple future tokens per step, and self-speculative verification converts those drafts into up to roughly 5x fewer generation steps.