Skip to content
NewA calmer, faster ZeroHour

Unifying Distributional Training for One-Step Visual Generation

arXiv cs.AI / cs.LG / cs.CLby Chi Zhang ()1 min read1 · 1 read
AI summary · grok-4.7Summaries by Vani AI

MGFlow unifies one-step visual training and beats FD-Loss on ImageNet and four-step FLUX.2.

The paper presents a unified distributional-training framework for one-step visual generation that separates distribution modeling from matching discrepancy using Wasserstein gradient flow. Under the framework, FD-Loss and Gaussian-kernel Drifting emerge as special cases, motivating MGFlow, which models features with Gaussian mixtures and pairs mass-constrained assignment with component updates to limit mode collapse. On ImageNet 256x256, MGFlow reports 1.45 FDr^6 on pMF-H and 1.64 on JiT-H, surpassing the FD-Loss baseline. Post-training FLUX.2 [klein] 4B into a one-step generator outperforms the original four-step model on GenEval and PickScore.

Key facts
  • Framework recovers FD-Loss and Gaussian-kernel Drifting as special cases.
  • MGFlow uses Gaussian mixtures with mass-constrained sample assignment.
  • ImageNet 256: 1.45 FDr^6 on pMF-H and 1.64 on JiT-H.
  • One-step FLUX.2 klein 4B beats the four-step model on GenEval and PickScore.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at arxiv.org

\emph{Distributional training} provides collective supervision for one-step visual generation by matching real and generated features in frozen representation spaces. We introduce \emph{a unified theoretical framework} that separates distribution modeling from matching discrepancy and connects global objectives to pointwise feature updates through Wasserstein gradient flow.

arXiv cs.AI / cs.LG / cs.CL

By Chi Zhang · arxiv.org

Read the full article at arXiv cs.AI / cs.LG / cs.CL

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.