Release Chronos-2 model on Hugging Face

[NielsRogge]

Hi @abdulfatir 🤗

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/2510.15821.
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.

It's fantastic to see that many Chronos family models (Chronos-Bolt, Chronos-T5) and the autogluon/chronos_datasets are already available on the Hugging Face Hub! This significantly improves their visibility and discoverability for the community.

I noticed that the main Chronos-2 model checkpoint is currently hosted on s3://autogluon/chronos-2. Would you be interested in also hosting the Chronos-2 model on https://huggingface.co/models? This would centralize access to your latest excellent work alongside your other Chronos models.

Hosting on Hugging Face will give it even more visibility and enable better discoverability. We can add tags in the model card so that people find the model easier, link it to the paper page, etc.

If you're down, here's a guide: https://huggingface.co/docs/hub/models-uploading. 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 for easy upload and usage by others. Alternatively, people can use hf_hub_download to download the model.

After uploading, we can also link the model to the paper page (read here) so people can discover your model.

You can also build a demo for your model on Spaces, and we can provide you a ZeroGPU grant, which gives you A100 GPUs for free.

Let me know if you're interested or need any guidance :)

Kind regards,

Niels

[abdulfatir]

@NielsRogge Thank you for reaching out and for the advice. We are currently working out some approvals with our open source team for the hugging face release. We hope to get this resolved soon.

[shchur]

Chronos-2 is now live on Hugging Face! https://huggingface.co/amazon/chronos-2

[abdulfatir]

@NielsRogge Like @shchur mentioned, Chronos-2 is live on HF but we need your kind help with a couple of things.

• When one clicks on the Use this model button, the transformers library shows up as the option and the code snippet is incorrect because the model only works with our inference package. How can we update this to point to this github repo with a custom snippet to avoid confusing the users?
• When one clicks on Deploy, there's an option to deploy on SageMaker but again the snippet won't work because it assumes that this model is an LLM which it is not. Chronos-2 is actually available on SageMaker JumpStart. How can we ensure that correct jumpstart usage snippet shows up there?

[NielsRogge]

Hi

• the first thing can be fixed by replacing library_name: transformers by a custom library you can add yourself and register on the hub. This way, you can also customize the code snippet of all models tagged with that library. Here's an example PR: Add YOLOv10 as library huggingface/huggingface.js#884 (you'll have to open a similar PR)
• the second one is a good question, will ask the team

[abdulfatir]

Thanks a lot @NielsRogge. For the first thing, I opened a PR here: huggingface/huggingface.js#1826

[abdulfatir]

@NielsRogge The first issue was resolved, I think. Still need to wait for it to go live. I'll wait for your update on the second issue.