Release MUSE framework and associated artifacts on Hugging Face

[NielsRogge]

Hi @Wayne-lc 🤗

Niels here from the open-source team at Hugging Face. I discovered your insightful work "Learning on the Job: An Experience-Driven Self-Evolving Agent for Long-Horizon Tasks" through Hugging Face's daily papers, and your paper got featured: https://huggingface.co/papers/2510.08002.
The paper page allows researchers to discuss your work and find related artifacts (such as models, datasets, or demos). You can also claim the paper as yours, which will be visible on your public Hugging Face profile, and add GitHub and project page URLs.

I noticed in the abstract that the code for the MUSE framework is planned to be released soon at https://github.com/KnowledgeXLab/MUSE. It would be fantastic if the code for the MUSE framework and any associated artifacts (such as specific implementations of the agent, or perhaps the accumulated "structured experience" if you plan to release it as a dataset) could be made available on the Hugging Face Hub once they are ready. This would greatly enhance the discoverability and visibility of your self-evolving agent framework.

Would you be interested in hosting the MUSE code and any related artifacts on the 🤗 Hub when you're ready for release? We can help you add tags to model/dataset cards so that people can easily find your work and link it directly to the paper page.

Uploading models (for agent implementations or fine-tuned checkpoints)

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 leverage the hf_hub_download one-liner to download a checkpoint from the Hub. You can also host your code directly in a Model or Dataset repository.

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 (for accumulated experience or similar)

If your framework generates or relies on new data (like the "structured experience" you mentioned) that you plan to release, it would be awesome to make it 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.

We can also provide you with a ZeroGPU grant to build a demo of your agent on Spaces, giving you access to A100 GPUs for free.

Please let me know if you're interested or need any help regarding this!

Cheers,

Niels
ML Engineer @ HF 🤗