Skip to content

Release HelloWorld model checkpoints and HelloWorldBench on Hugging Face#1

Description

@NielsRogge

Hello @AlayaLab 馃

Niels here from the open-source team at Hugging Face. I discovered your work through Hugging Face Daily Papers: https://huggingface.co/papers/2608.05070
The paper page lets people discuss your paper and discover related artifacts (such as your model checkpoints and benchmark dataset). You can also claim the paper as yours on HF to display it on your public profile, as well as link your GitHub and project page URLs.

I noticed in your repository that the code and benchmark dataset (HelloWorldBench) are marked as coming soon. When you are ready to release them, it would be great to host the model checkpoints on https://huggingface.co/models and the benchmark dataset on https://huggingface.co/datasets to improve their discoverability and visibility in the open-source AI community.

Uploading models

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

You can leverage the PyTorchModelHubMixin class which adds from_pretrained and push_to_hub to custom nn.Module classes, or use the hf_hub_download one-liner to download checkpoints from the hub.

We encourage pushing each model checkpoint to a separate model repository so that download statistics work properly. We can then also link the checkpoints directly to the paper page.

Uploading dataset

Would be awesome to make HelloWorldBench available on Hugging Face, so that people can easily load it:

from datasets import load_dataset

dataset = load_dataset("AlayaLab/HelloWorldBench")

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

Additionally, our dataset viewer will allow users to explore and preview the benchmark data directly in their browser.

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

Cheers,

Niels
ML Engineer @ HF 馃

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions