@inproceedings{73971c84b0ef401e92f11177407783cb,
title = "DynaMo: Accelerating Language Model Inference with Dynamic Multi-Token Sampling",
abstract = "Traditional language models operate autoregressively, i.e., they predict one token at a time. Rapid explosion in model sizes has resulted in high inference times. In this work, we propose DynaMo, a suite of multi-token prediction language models that reduce net inference times. Our models dynamically predict multiple tokens based on their confidence in the predicted joint probability distribution. We propose a lightweight technique to train these models, leveraging the weights of traditional autoregressive counterparts. Moreover, we propose novel ways to enhance the estimated joint probability to improve text generation quality, namely co-occurrence weighted masking and adaptive thresholding. We also propose systematic qualitative and quantitative methods to rigorously test the quality of generated text for non-autoregressive generation. One of the models in our suite, DynaMo-7.3BT3, achieves same-quality generated text as the baseline (Pythia-6.9B) while achieving 2.57× speed-up with only 5.87% and 2.67% parameter and training time overheads, respectively.",
author = "Shikhar Tuli and Lin, {Chi Heng} and Hsu, {Yen Chang} and Jha, {Niraj K.} and Yilin Shen and Hongxia Jin",
note = "Publisher Copyright: {\textcopyright}2024 Association for Computational Linguistics.; 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2024 ; Conference date: 16-06-2024 Through 21-06-2024",
year = "2024",
language = "English (US)",
series = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2024",
publisher = "Association for Computational Linguistics (ACL)",
pages = "3322--3345",
editor = "Kevin Duh and Helena Gomez and Steven Bethard",
booktitle = "Long Papers",
}