Doubiiu / DynamiCrafter

[ECCV 2024, Oral] DynamiCrafter: Animating Open-domain Images with Video Diffusion Priors
https://doubiiu.github.io/projects/DynamiCrafter/
Apache License 2.0
2.57k stars 205 forks source link

Fix download stats on Hugging Face, improve discoverability #108

Open NielsRogge opened 3 months ago

NielsRogge commented 3 months ago

Hi DynamicCrafter team,

Congrats on this work! It's great to see the checkpoints being released on HF 🤗

Download stats

I see that currently download stats aren't working for your model. This is due to the model repositories not containing a config.json or config.yaml file. See here for more info: https://huggingface.co/docs/hub/models-download-stats.

There are a few options here to make them work:

1) either we can open a PR on the huggingface.js repository to view "DynamiCrafter" as a library, for which we add dedicated support. See this PR as an example: https://github.com/huggingface/huggingface.js/pull/784 2) one could leverage the PyTorchModelHubMixin which is a minimal class which can be added to any nn.Module, to add from_pretrained and push_to_hub capabilities. This ensures the safer safetensors format is leveraged.

Improve discoverability

Besides making download stats work, we could also help in improving the discoverability of your models on the hub. We can for instance add appropriate tags to the model repos (READMEs) such as text-to-video which allow people to find your models.

Let me know what you prefer, and whether you need any help regarding this!

Cheers,

Niels ML Engineer @ HF 🤗

NielsRogge commented 3 months ago

Opened https://github.com/huggingface/huggingface.js/pull/804 to address this