Release DicFace models on Hugging Face

[NielsRogge]

Hi @SiyuZhu 🤗

I'm Niels, and I work as part of the open-source team at Hugging Face. I came across your work, DicFace, on arXiv and was impressed
by its state-of-the-art performance in video face restoration. I noticed your paper is also featured on our Hugging Face Papers page:
https://huggingface.co/papers/2506.13355.

I'm reaching out to see if you'd be interested in hosting your pre-trained models on the Hugging Face Hub. This would greatly increase
the visibility and accessibility of your work, allowing other researchers to easily utilize your models for their own projects. We can also help
you add detailed metadata to your model cards, making them more discoverable.

You can find a guide on uploading models here: https://huggingface.co/docs/hub/models-uploading.
For PyTorch models, the PyTorchModelHubMixin simplifies the upload process:
https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin

We can also link your models to your Hugging Face Papers page for improved discoverability:
https://huggingface.co/docs/hub/en/model-cards#linking-a-paper

Please let me know if you're interested or have any questions.

Kind regards,

Niels
ML Engineer @ Hugging Face 🤗

[NinoNeumann]

Thank you so much for your kind words about DicFace and for reaching out! We will upload the pre-trained models to Hugging Face very soon.

[NinoNeumann]

Hi Niels,

Thanks for reaching out! We’ve already uploaded the pretrained weights to Hugging Face Hub. Here’s the link: https://huggingface.co/fudan-generative-ai/DicFace_model.

Happy to collaborate on metadata or linking to the Papers page. Let us know if you need anything else!

Best,
Hanlin Shang

[NielsRogge]

Awesome! I've opened a PR to link it to the paper page and add more metadata: https://huggingface.co/Alibaba-NLP/WebDancer-32B/discussions/1