ProPD: Dynamic Token Tree Pruning and Generation for LLM Parallel Decoding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhong, Shuzhang, Yang, Zebin, Li, Meng, Gong, Ruihao, Wang, Runsheng, Huang, Ru
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913238428418048
author Zhong, Shuzhang
Yang, Zebin
Li, Meng
Gong, Ruihao
Wang, Runsheng
Huang, Ru
author_facet Zhong, Shuzhang
Yang, Zebin
Li, Meng
Gong, Ruihao
Wang, Runsheng
Huang, Ru
contents Recent advancements in generative large language models (LLMs) have significantly boosted the performance in natural language processing tasks. However, their efficiency is hampered by the inherent limitations in autoregressive token generation. While parallel decoding with token tree verification, e.g., Medusa, has been proposed to improve decoding parallelism and efficiency, it often struggles with maintaining contextual relationships due to its independent token prediction approach and incurs significant verification overhead, especially with large tree sizes and batch processing. In this paper, we propose ProPD, an efficient LLM parallel decoding framework based on dynamic token tree pruning and generation. ProPD features an advanced early pruning mechanism to efficiently eliminate unpromising token sequences to improve verification efficiency. Additionally, it introduces a dynamic token tree generation algorithm to balance the computation and parallelism of the verification phase in real-time and maximize the overall efficiency across different batch sizes, sequence lengths, and tasks, etc. We verify ProPD across a diverse set of datasets, LLMs, and batch sizes and demonstrate ProPD consistently outperforms existing decoding algorithms by 1.1-3.2x.
format Preprint
id arxiv_https___arxiv_org_abs_2402_13485
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ProPD: Dynamic Token Tree Pruning and Generation for LLM Parallel Decoding
Zhong, Shuzhang
Yang, Zebin
Li, Meng
Gong, Ruihao
Wang, Runsheng
Huang, Ru
Machine Learning
Computation and Language
Recent advancements in generative large language models (LLMs) have significantly boosted the performance in natural language processing tasks. However, their efficiency is hampered by the inherent limitations in autoregressive token generation. While parallel decoding with token tree verification, e.g., Medusa, has been proposed to improve decoding parallelism and efficiency, it often struggles with maintaining contextual relationships due to its independent token prediction approach and incurs significant verification overhead, especially with large tree sizes and batch processing. In this paper, we propose ProPD, an efficient LLM parallel decoding framework based on dynamic token tree pruning and generation. ProPD features an advanced early pruning mechanism to efficiently eliminate unpromising token sequences to improve verification efficiency. Additionally, it introduces a dynamic token tree generation algorithm to balance the computation and parallelism of the verification phase in real-time and maximize the overall efficiency across different batch sizes, sequence lengths, and tasks, etc. We verify ProPD across a diverse set of datasets, LLMs, and batch sizes and demonstrate ProPD consistently outperforms existing decoding algorithms by 1.1-3.2x.
title ProPD: Dynamic Token Tree Pruning and Generation for LLM Parallel Decoding
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2402.13485