Host model on HF
Hi,
Niels here from the open-source team at Hugging Face. I discovered your work through the paper page: https://huggingface.co/papers/2408.05205 (feel free to claim the paper so that it appears under your HF account!).
To improve the discoverability of your work, it would be great to make the models available on the 🤗 hub (along with tags like video-super-resolution which allows people to find it).
See here for a guide: https://huggingface.co/docs/hub/models-uploading. In this case, the PyTorchModelHubMixin might be useful given it's a custom PyTorch model.
Let me know if you need any help regarding this!
Cheers,
Niels ML Engineer @ HF 🤗
Thanks for the suggestion. We are working on the hugging demo.
Hi @NielsRogge ,
I have made a gradio demo and uploaded to huggingface at https://huggingface.co/spaces/rcfeng/KEEP. But I found it much slower than running locally. Do you have any suggestion?
It seems to be running on CPU. Do you require a GPU?
Sure, that would be great for more people to try it on their own videos.
You can apply for a free grant for ZeroGPU here: https://huggingface.co/docs/hub/en/spaces-gpus#community-gpu-grants
Thanks. I've opened a discussion at https://huggingface.co/spaces/rcfeng/KEEP/discussions/1.
Great! Btw I noticed all models are on Github: https://github.com/jnjaby/KEEP/releases/tag/v1.0.0
Would be nice to make them available in one or multiple model repositories on the hub, we could link them to your paper: https://huggingface.co/papers/2408.05205
Also, the synthetic dataset is on Google Drive I see, would be great to enhance discoverability/ease of access on the hub! Happy to help. Here's a guide: https://huggingface.co/docs/datasets/en/image_dataset