Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Appearance settings

WeichenFan/UAE

Open more actions menu

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

The Prism Hypothesis: Harmonizing Semantic and Pixel Representations via Unified Autoencoding

S-Lab, Nanyang Technological University1      SenseTime Research 2
Corresponding Author.


🔥 More coming soon!

Installation

conda create -n uae python=3.10 -y
conda activate uae
pip install uv

uv pip install torch==2.2.0 torchvision==0.17.0 torchaudio --index-url https://download.pytorch.org/whl/cu121

uv pip install timm==0.9.16 accelerate==0.23.0 torchdiffeq==0.2.5 wandb
uv pip install "numpy<2" transformers einops omegaconf
uv pip install torchmetrics

Quick evaluation

python eval_uae.py \
  --config unified_ae/configs/stage1_infer.yaml \
  --checkpoint PATH_TO_WEIGHTS \
  --imagenet-path PATH_TO_IMAGENET \
  --coco-path PATH_TO_COCO \
  --batch-size 16 \
  --num-workers 8 \
  --image-size 256 \
  --freq-ratio 1.0 \
  --log-file logs/uae_eval_metrics.txt

Expected Results:

ImageNet: PSNR=29.588 dB | SSIM=0.8789 | rFID=0.193
MS-COCO: PSNR=29.484 dB | SSIM=0.8846 | rFID=0.157

BibTex

@misc{fan2025uae,
      title={The Prism Hypothesis: Harmonizing Semantic and Pixel Representations via Unified Autoencoding}, 
      author={Weichen Fan and Haiwen Diao and Quan Wang and Dahua Lin and Ziwei Liu},
      year={2025},
      eprint={2512.19693},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2512.19693}, 
}

✨ Star History

Star History Chart

About

Official repo for UAE

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

Morty Proxy This is a proxified and sanitized view of the page, visit original site.