Hi @JackAILab 🤗
Niels here from the open-source team at Hugging Face. I discovered your work through Papers with Code and your paper page on Hugging Face: https://huggingface.co/papers/2606.21661.
First of all, congratulations on the project! It's awesome to see the UnityShotsBench dataset already hosted on the Hugging Face hub under the KlingTeam organization.
I saw in your GitHub roadmap that the model checkpoints for UnityShots (T2V/I2V/R2V weights) are coming soon. Would you also be interested in hosting these model checkpoints directly on Hugging Face at https://huggingface.co/models?
Hosting on Hugging Face will improve the discoverability of your work, and we can link the model repositories directly to your Arxiv paper page so that users can easily find them.
If you are interested, we have a quick guide here. If you are using PyTorch, you can also leverage the PyTorchModelHubMixin class which adds from_pretrained and push_to_hub to your custom model class, allowing people to download and use the models right away. Alternatively, users can easily fetch weights using hf_hub_download.
Additionally, we would love to help you build a demo for your model on Spaces and can provide you with a ZeroGPU grant to run your demo Space for free.
Let me know if you're interested or if you need any guidance when the weights are ready for release!
Kind regards,
Niels
Hi @JackAILab 🤗
Niels here from the open-source team at Hugging Face. I discovered your work through Papers with Code and your paper page on Hugging Face: https://huggingface.co/papers/2606.21661.
First of all, congratulations on the project! It's awesome to see the
UnityShotsBenchdataset already hosted on the Hugging Face hub under the KlingTeam organization.I saw in your GitHub roadmap that the model checkpoints for UnityShots (T2V/I2V/R2V weights) are coming soon. Would you also be interested in hosting these model checkpoints directly on Hugging Face at https://huggingface.co/models?
Hosting on Hugging Face will improve the discoverability of your work, and we can link the model repositories directly to your Arxiv paper page so that users can easily find them.
If you are interested, we have a quick guide here. If you are using PyTorch, you can also leverage the PyTorchModelHubMixin class which adds
from_pretrainedandpush_to_hubto your custom model class, allowing people to download and use the models right away. Alternatively, users can easily fetch weights using hf_hub_download.Additionally, we would love to help you build a demo for your model on Spaces and can provide you with a ZeroGPU grant to run your demo Space for free.
Let me know if you're interested or if you need any guidance when the weights are ready for release!
Kind regards,
Niels