AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

LayerSync: Self-aligning Intermediate Layers

Yasaman Haghighi, Bastien Van Delft, Mariam Hassan, Alexandre Alahi

ICLR 2026 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers · 8 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add LayerSync self-alignment of shallow to deep layersSiT-XL/2 · ImageNet 256x256 (800 epochs, no CFG)FID: 8.99 -> 6.87Improved
add LayerSync instead of Dispersive lossSiT-XL/2 · ImageNet 256x256 (80 epochs)FID: 15.95 -> 11.24Improved
add LayerSyncSiT-B/2 · ImageNet 256x256 (80 epochs)FID: 36.19 -> 30.00Improved
add LayerSyncSiT-XL/2 with SDE sampler · ImageNet 256x256FID: 8.3 (1400 ep) -> 8.29 (160 ep)Improved
add LayerSyncMTG-Jamendo audio SiT-XL · MTG-JamendoFAD (CLAP): 0.251 -> 0.199Improved
add LayerSyncMDM · HumanML3DFID: 0.5206 -> 0.4801Improved
add LayerSync to REPASiT-XL/2 + REPA · ImageNet 256x256 (200k iterations)FID: 7.88 -> 7.01Improved
use random layer pairings instead of the heuristic layer selection (ablation)LayerSync SiT-XL · ImageNet 256x256FID: 12.24 (std 0.8) vs baseline 17.98No clear effect
vary lambda from 0.1 to 0.7 (ablation)LayerSync SiT-B/2 · ImageNet 256x256FID: 31.02 to 31.63No clear effect
add LayerSync, compare 160-epoch LayerSync model vs 1400-epoch baseline (ablation)SiT-XL/2 · Tiny ImageNet linear probingavg accuracy: +32.4% avg accuracyImproved
prune layers of a LayerSync-trained model (ablation)SiT-XL/2 · layer removalgeneration qualityHurt
use LayerSync* with CFG scheduling vs REPA*SiT-XL/2 · ImageNet 256x256 (with CFG, 800 epochs)FID: 1.49 vs REPA* 1.42Mixed

Later papers that compared with this method

1 later accepted paper reports + LayerSync as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy