4 papers
Accelerating LLM Inference with Lossless Speculative Decoding Algorithms for Heterogeneous Vocabularies
Nadav Timor, Jonathan Mamou, Daniel Korat +5
Accelerating the inference of large language models (LLMs) is a critical challenge in generative AI. Speculative decoding (SD) methods offer substantial efficiency gains by generat…
Out-of-Vocabulary Sampling Boosts Speculative Decoding
Nadav Timor, Jonathan Mamou, Oren Pereg +2
Speculative decoding relies on fast and accurate drafters. Recent state-of-the-art language models employ larger and larger vocabularies, which significantly slows down drafters. O…
Distributed Speculative Inference (DSI): Speculation Parallelism for Provably Faster Lossless Language Model Inference
Nadav Timor, Jonathan Mamou, Daniel Korat +6
This paper introduces distributed speculative inference (DSI), a novel inference algorithm that is provably faster than speculative inference (SI) [leviathan2023, chen2023, miao202…
Dynamic Speculation Lookahead Accelerates Speculative Decoding of Large Language Models
Jonathan Mamou, Oren Pereg, Daniel Korat +4
Speculative decoding is commonly used for reducing the inference latency of large language models. Its effectiveness depends highly on the speculation lookahead (SL)-the number of…