Introduction
SANA combines deep latent compression, linear attention and efficient sampling for high-resolution synthesis, connecting model design with computation and local production conditions.
View full image ↗Figure 1: SANA samples and the paper’s latency comparisons. Source: v3; the image version is distinct from the initial submission date.
Xie, Enze; Chen, Junsong; Chen, Junyu; Cai, Han; Tang, Haotian; Lin, Yujun; Zhang, Zhekai; Li, Muyang; Zhu, Ligeng; Lu, Yao; Han, Song · 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 · SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers ↗RESEARCH ACCOUNT
SANA combines deep latent compression, linear attention and efficient sampling for high-resolution synthesis, connecting model design with computation and local production conditions.
ANTI-AI ARCHIVE · Revised 2026-10-03
What the paper investigates
Deeper compression reduces latent-token counts, and linear attention addresses computation at high resolutions. Text encoding, training and sampling choices form part of the design. Reported efficiency depends on the hardware, model size and evaluation configuration.
Section sources: SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers
Its place in generative-art history
Editorial interpretation: repeated generation on smaller devices can affect who can make, revise and retain images. SANA offers technical evidence about those material conditions, without suggesting that lower computation removes data or rights questions.
Section sources: SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers
Reading limits and versions
Throughput ratios are not generalised to every device; local runtime performance was not tested.
Section sources: SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers
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
SANA: compare methods, control or evaluation conditions with DPM-Solver.
SANA: compare methods, control or evaluation conditions with A new model ecosystem.
SANA: compare methods, control or evaluation conditions with Stable Diffusion 3.5.
Dates & version record
Date displayed for this node: 2024-10-14 · Historical date recorded for the source: 2024-10-14
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 v3; 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.
SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers
https://arxiv.org/abs/2410.10629
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 #293 · 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. · 028
Date as recorded: 2024-10-14
SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers ↗
SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers
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. SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers. Art-history node #293. https://salondesrefuses.cn/en/art-history/351
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.