Introduction
MeanFlow characterises generative flows through interval-average rather than instantaneous velocity and trains one-step models from scratch, extending the history of few-step generation.
View full image ↗Figure 1: MeanFlow one-step samples and the paper’s quality comparison. Source: v1; the image version is distinct from the initial submission date.
Geng, Zhengyang; Deng, Mingyang; Bai, Xingjian; Kolter, J. Zico; He, Kaiming · original paper / arXiv · Rights in the paper and depicted works remain with their holders. Research quotation does not establish an open licence; republication rights await independent review.
Source · Mean Flows for One-step Generative Modeling ↗RESEARCH ACCOUNT
MeanFlow characterises generative flows through interval-average rather than instantaneous velocity and trains one-step models from scratch, extending the history of few-step generation.
ANTI-AI ARCHIVE · Revised 2026-10-03
What the paper investigates
An identity linking average and instantaneous velocities guides the training objective. The method does not require a pretrained teacher or distillation. ImageNet experiments examine quality when generation uses a single network evaluation.
Section sources: Mean Flows for One-step Generative Modeling
Its place in generative-art history
Editorial interpretation: this adds a route to fast generation distinct from distillation. Response time matters to interactive production, while adoption into stable creative workflows still requires evidence about particular tools and practices.
Section sources: Mean Flows for One-step Generative Modeling
Reading limits and versions
One network evaluation is an experimental measure; a complete application can still incur encoding, decoding and other computational costs.
Section sources: Mean Flows for One-step Generative Modeling
Sources for this account
Read the arXiv abstract, authors and version history, plus the selected figure/page on PDF page 1; proofs and full experiments were not independently audited.
Read the arXiv abstract, authors and version history, plus the selected figure/page on PDF page 1; proofs and full experiments were not independently audited. The account and translation are AI-assisted, pending independent human review. Section references identify evidence without claiming independent verification of every historical statement.
Continue with a comparative question
MeanFlow: compare methods, control or evaluation conditions with Consistency Models.
MeanFlow: compare methods, control or evaluation conditions with Flow Matching.
MeanFlow: compare methods, control or evaluation conditions with Latent Consistency Models.
MeanFlow: compare methods, control or evaluation conditions with Stable Diffusion 3 / Scaling Rectified Flow Transformers for High-Resolution Image Synthesis.
Dates & version record
Date displayed for this node: 2025-05-19 · Historical date recorded for the source: 2025-05-19
The timeline uses the initial arXiv submission, distinct from conference publication, model release and the archive addition on 3 October 2026. The account and image refer to v1; its date is recorded in the source history.
These dates refer to the historical event or recorded version, not this page’s publication date. The original date precision and unresolved questions are retained.
Original sources & further reading
These links lead to the cited paper, article, institution or conference page. External texts retain their source languages.
Mean Flows for One-step Generative Modeling
https://arxiv.org/abs/2505.13447
Current cited URL
Read the arXiv abstract, authors and version history, plus the selected figure/page on PDF page 1; proofs and full experiments were not independently audited.
Provenance, translation & verification
Archive node #326 · Initial research-paper submission; methods, evaluation and production conditions
Research materials & supplement references · 1
Recent research papers: additions and reconciled sources since 2022 · Read the arXiv abstract, authors and version history, plus the selected figure/page on PDF page 1; proofs and full experiments were not independently audited. · 032
Date as recorded: 2025-05-19
Mean Flows for One-step Generative Modeling ↗
Mean Flows for One-step Generative Modeling
Archive account based on the listed research materials, not a full translation of the linked work. AI-assisted translation; independent human review pending.
Official material was read within the stated scope; see the source note for reading limits and outstanding checks.
Cite this node
ANTI-AI ARCHIVE. Mean Flows for One-step Generative Modeling. Art-history node #326. https://salondesrefuses.cn/en/art-history/355
For specific historical claims, also cite the original sources above and include your access date. This account is not a full translation of the linked work.
Adjacent nodes follow chronological order; adjacency does not establish direct influence or causation.