Partition Generative Modeling: Masked Modeling Without Masks

Justin Deschenaux, Lan Tran, Caglar Gulcehre

Keywords: masked generative modeling, discrete diffusion, masked diffusion language modeling, diffusion language modeling

TL;DR: We show that it is possible to train masked generative models without using MASK tokens, resulting in efficiency gains at inference.

Abstract:

Masked generative models (MGMs) can generate tokens in parallel and in any order, unlike autoregressive models (ARMs), which decode one token at a time, left-to-right. However, MGMs process the full-length sequence at every sampling step, including \mask tokens that carry no information. In contrast, ARMs process only the previously generated tokens.
We introduce ``Partition Generative Models'' (PGMs), which replace masking with partitioning. Tokens are split into two groups that cannot attend to each other, and the model learns to predict each group conditioned on the other, eliminating mask tokens entirely. Because the groups do not interact, PGMs can process only the clean tokens during sampling, like ARMs, while retaining parallel, any-order generation, like MGMs.
On OpenWebText, PGMs achieve $5-5.5\times$ higher throughput than MDLM while producing samples with lower Generative Perplexity. On ImageNet, PGMs reach comparable FID to MaskGIT with a $7.5\times$ throughput improvement. With twice as many steps, the FID improves to 4.56 while remaining $3.9\times$ faster than MGMs. Finally, PGMs remain compatible with existing MGM samplers and distillation methods.

Primary Area: generative models

Submission Number: 7931