Release Nemotron Nano V2 VL models and datasets on Hugging Face

[NielsRogge]

Hi @taesiri 🤗

Niels here from 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/2511.03929. The paper page lets people discuss about your paper and lets them find artifacts about it (your models, datasets or demo 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.

I saw in your paper that you are releasing model checkpoints for Nemotron Nano V2 VL (in BF16, FP8, and FP4 formats) and also sharing large parts of your datasets, including Nemotron VLM Dataset V2 and NVPDFTex. This is fantastic!

I couldn't find direct links to these new artifacts (models and datasets) in the provided materials, and the GitHub link for the code resulted in a 404. It would be great to make these new checkpoints and datasets easily discoverable and accessible on the 🤗 Hub, to improve their visibility. We can add tags so that people find them when filtering https://huggingface.co/models and https://huggingface.co/datasets.

Could you please share the specific Hugging Face repository links for the new Nemotron Nano V2 VL models and the Nemotron VLM Dataset V2 and NVPDFTex datasets?

Uploading models

See here for a guide: https://huggingface.co/docs/hub/models-uploading.

In this case, we could leverage the PyTorchModelHubMixin class which adds from_pretrained and push_to_hub to any custom nn.Module. Alternatively, one can leverages the hf_hub_download one-liner to download a checkpoint from the hub.

We encourage researchers to push each model checkpoint to a separate model repository, so that things like download stats also work. We can then also link the checkpoints to the paper page.

Uploading dataset

Would be awesome to make the dataset available on 🤗 , so that people can do:

from datasets import load_dataset

dataset = load_dataset("your-hf-org-or-username/your-dataset")

See here for a guide: https://huggingface.co/docs/datasets/loading.

Besides that, there's the dataset viewer which allows people to quickly explore the first few rows of the data in the browser.

Let me know if you're interested/need any help regarding this!

Cheers,

Niels ML Engineer @ HF 🤗

[ayushdg]

Hey @NielsRogge. Looking through the pdf version of the paper I do see updated links to the models and v2 dataset. The link to the code in Curator https://github.com/NVIDIA-NeMo/Curator/tree/experimental/experimental/nvpdftex seems to no longer return a 404.

Are there spots where you still see broken links or is this good to close?

[sithape2025]

Hey @NielsRogge , following up to see if there are any further updates on the thread.

[NielsRogge]

It looks like https://huggingface.co/papers/2511.03929 has 3 linked models, so that's great. Not sure whether the dataset will be released too.

[ayushdg]

Thanks for confirming. Looks like the pdf/arxiv paper does link to the dataset here but the dataset card is missing the citation that links it to the papers page.

I've asked the uploader to take a look and update the card and the issue is probably more relevant to https://huggingface.co/datasets/nvidia/Nemotron-VLM-Dataset-v2/discussions/6 rather than this repo. Closing in favor of tracking this in the linked discussion there. Feel free to re-open if you have any questions or comments.

[NielsRogge]

Oh yes thanks! Would be useful to merge that PR so that the dataset shows up on the paper page.