Discrete Flow Maps

TL;DR

Discrete Flow Maps introduce a novel framework for sequence generation that compresses trajectories into single-step mappings, aligning with the geometry of discrete data to improve language modeling efficiency.

Contribution

The paper presents Discrete Flow Maps, a new approach that adapts flow models to discrete data, overcoming geometric and training challenges for improved language sequence generation.

Findings

Outperforms previous state-of-the-art in discrete flow modeling.

Reconciles trajectory compression with the geometry of the probability simplex.

Enables single-pass sequence generation from noise.

Abstract

The sequential nature of autoregressive next-token prediction imposes a fundamental speed limit on large language models. While continuous flow models offer a path to parallel generation, they traditionally demand expensive iterative integration. Flow Maps bypass this bottleneck by compressing generative trajectories into single-step mappings, theoretically enabling the generation of full text sequences from noise in a single forward pass. However, standard formulations rely on Euclidean regression losses that are geometrically ill-suited for discrete data. In this work, we resolve this conflict with Discrete Flow Maps, a framework that reconciles trajectory compression with the geometry of the probability simplex. We recast standard flow map training for the discrete domain, aligning the training dynamics with the discrete nature of language. Empirically, this strict geometric alignment allows our method to surpass previous state-of-the-art results in discrete flow modeling.