← Monday, September 28, 2026Models & trainingPaper

FuseReg improves image generation by training across encoder-layer subsets

huggingface.co by Hongyang Du et al. Editor score 8/10

FuseReg trains a decoder on random subsets of encoder layers; on ImageNet-256 it reportedly improves reconstruction flexibility and cuts unguided gFID by 27% through decoder replacement alone. The result suggests a practical way to separate representation choices from generator retraining.

What

FuseReg trains decoders on random subsets of pretrained encoder layers rather than fixed layer fusions.

Method

On ImageNet-256 with DINOv3-L, one decoder handles full, sparse, and single-layer fusions without retraining and achieves higher PSNR than fixed-fusion decoders.

Why it matters

Decoder replacement alone reduced unguided gFID by 27% with the RAEv2 DiT-XL generator unchanged.

Read on huggingface.co → Code · 2 stars

Headline, summary, brief and diagram are written by a model from the source's own text; the link, byline and numbers come from the source. Read the original before relying on a claim.