Hi @MR-Sherif 🤗
I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Hugging Face's paper pages as yours was recently featured: https://huggingface.co/papers/2603.18101.
The paper page lets people discuss 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, and add Github and project page URLs.
I saw on your GitHub repository that the implementation of TOGA is coming soon! Would you like to host the pre-trained checkpoints (the distilled key-value caches and adapter weights) for the various benchmarks on https://huggingface.co/models?
Hosting on Hugging Face will give your work more visibility and enable better discoverability within the community. We can add metadata tags (like zero-shot-image-classification) to the model cards so that people find the models easier and link them directly to the paper page.
If you're down, I'm leaving a guide here. If it's a custom PyTorch model, you can use the PyTorchModelHubMixin class which adds from_pretrained and push_to_hub to the model, allowing people to download and use your work right away. Alternatively, researchers can use hf_hub_download to fetch specific cache files.
After they are uploaded, we can also link the models to the paper page so people can discover your work more easily.
Let me know if you're interested or need any guidance!
Kind regards,
Niels
Hi @MR-Sherif 🤗
I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Hugging Face's paper pages as yours was recently featured: https://huggingface.co/papers/2603.18101.
The paper page lets people discuss 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, and add Github and project page URLs.
I saw on your GitHub repository that the implementation of TOGA is coming soon! Would you like to host the pre-trained checkpoints (the distilled key-value caches and adapter weights) for the various benchmarks on https://huggingface.co/models?
Hosting on Hugging Face will give your work more visibility and enable better discoverability within the community. We can add metadata tags (like
zero-shot-image-classification) to the model cards so that people find the models easier and link them directly to the paper page.If you're down, I'm leaving a guide here. If it's a custom PyTorch model, you can use the PyTorchModelHubMixin class which adds
from_pretrainedandpush_to_hubto the model, allowing people to download and use your work right away. Alternatively, researchers can use hf_hub_download to fetch specific cache files.After they are uploaded, we can also link the models to the paper page so people can discover your work more easily.
Let me know if you're interested or need any guidance!
Kind regards,
Niels