Release TiPToP and manipulation benchmark on Hugging Face

[NielsRogge]

Hi @williamshen-nz 🤗

I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work on Arxiv and was wondering whether you would like to submit it to hf.co/papers to improve its discoverability. If you are one of the authors, you can submit it at https://huggingface.co/papers/submit.

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 on your GitHub repository that the code for TiPToP is "coming soon". It'd be great to make the system and the benchmark of 28 tabletop manipulation tasks available on the 🤗 hub, to improve their discoverability/visibility. We have a growing robotics community, and we can add tags so that people find them when filtering https://huggingface.co/models and https://huggingface.co/datasets.

Uploading models/code

You can host the system code and any configuration files in a model repository. See here for a guide: https://huggingface.co/docs/hub/models-uploading. Alternatively, one can leverages the hf_hub_download one-liner to download any necessary components.

Uploading dataset

It would also be awesome to make the evaluation benchmark available on 🤗 , so that people can easily explore the task configurations:

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 🤗