References & Citations
Computer Science > Computation and Language
Title: Accelerating Production LLMs with Combined Token/Embedding Speculators
(Submitted on 29 Apr 2024)
Abstract: This technical report describes the design and training of novel speculative decoding draft models, for accelerating the inference speeds of large language models in a production environment. By conditioning draft predictions on both context vectors and sampled tokens, we can train our speculators to efficiently predict high-quality n-grams, which the base model then accepts or rejects. This allows us to effectively predict multiple tokens per inference forward pass, accelerating wall-clock inference speeds of highly optimized base model implementations by a factor of 2-3x. We explore these initial results and describe next steps for further improvements.
Submission history
From: Davis Wertheimer [view email][v1] Mon, 29 Apr 2024 21:59:07 GMT (2397kb,D)
Link back to: arXiv, form interface, contact.