Release AlphaFlow models on Hugging Face #1
Description
Hi @huijieZH 🤗
I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Hugging Face's daily papers as yours got featured: https://huggingface.co/papers/2510.20771.
The paper page lets people discuss about your paper and lets them find artifacts about it (your models for instance), you can also claim the paper as yours which will show up on your public profile at HF, add Github and project page URLs.
It's great to see that you've already made the checkpoints (MeanFlow-B/2, α-Flow-B/2, etc.) available for download directly from your snap-research/alphaflow repository on Hugging Face!
To maximize their discoverability and visibility, we'd love to help you publish these as individual model repositories on the 🤗 Hub (https://huggingface.co/models). This approach offers several benefits:
- Enhanced Discoverability: Models become easily searchable and filterable on the Hugging Face Models page.
- Rich Metadata: You can create comprehensive model cards with detailed descriptions, usage examples, performance metrics, and appropriate tags.
- Usage Statistics: Track downloads and other engagement metrics for each specific model.
- Seamless Integration: Enable direct loading with
from_pretrainedand easy pushing withpush_to_hubfunctionalities. - Linking to Paper Page: We can directly link each model card to your paper page, making it super easy for researchers to find your models from the paper.
If you're interested in doing this, here's a guide: https://huggingface.co/docs/hub/models-uploading.
For custom PyTorch models, the PyTorchModelHubMixin class adds from_pretrained and push_to_hub, making it straightforward to upload and for others to use your models.
Alternatively, users can always use hf_hub_download to get specific files.
You can also build a demo for your model on Spaces, and we can provide you a ZeroGPU grant for free A100 GPUs.
Let me know if you're interested or need any guidance!
Kind regards,
Niels