Verify evals on Papers with Code

Hi,

Niels here from the open-source team at Hugging Face. Congratulations on your work!

I've made the [paper](https://paperswithcode.co/paper/2407.17331) and [38 paper-native evaluations](https://paperswithcode.co/paper/2407.17331#results) available on Papers with Code.

The paper has results on [Image Classification](https://paperswithcode.co/tasks/image-classification), [Image Understanding](https://paperswithcode.co/tasks/image-understanding), and [Text classification](https://paperswithcode.co/tasks/text-classification) task pages.

The MLCD ViT-L/14 (linear probe) results currently rank second on [Caltech101](https://paperswithcode.co/benchmark/caltech101?task=image-classification&eval=7330) and [EuroSAT](https://paperswithcode.co/benchmark/eurosat?task=image-classification&eval=7332).

The MLCD ViT-L/14 (linear probe) results currently rank third on [CLEVR Count](https://paperswithcode.co/benchmark/clevr-count?task=image-understanding&eval=7334) and [DTD](https://paperswithcode.co/benchmark/dtd?task=image-classification&eval=7328).

Would it be possible to verify these results and let me know if any score, model name, benchmark protocol, or openness metadata should be corrected?

You can also edit the task, methods, project page, and GitHub URL directly from the paper page using your Hugging Face account.

If you'd like to showcase the results in your repository README, you can copy these live leaderboard badges (or use the “Copy PwC badge” button in the Results section):

[](https://paperswithcode.co/api/v1/papers/2407.17331/leaderboard-badge-link?eval=7330)
[](https://paperswithcode.co/api/v1/papers/2407.17331/leaderboard-badge-link?eval=7332)
[](https://paperswithcode.co/api/v1/papers/2407.17331/leaderboard-badge-link?eval=7334)
[](https://paperswithcode.co/api/v1/papers/2407.17331/leaderboard-badge-link?eval=7328)
[](https://paperswithcode.co/api/v1/papers/2407.17331/leaderboard-badge-link?eval=7325)

Kind regards,

Niels

Start with paper 2407.17331 and its 38 paper-native evaluations on Papers with Code. Compare the listed scores, model name, benchmark protocols, openness metadata, and linked task pages with the repository's paper materials. Done means any discrepancies are identified and the Papers with Code entries or project documentation are corrected.

Written by the indexing model from the issue text.