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/2408.00714) and [5 paper-native evaluations](https://paperswithcode.co/paper/2408.00714#results) available on Papers with Code.

The paper has results on [Video segmentation](https://paperswithcode.co/tasks/video-segmentation) and [Image segmentation](https://paperswithcode.co/tasks/image-segmentation) task pages.

The SAM 2 (full data mix) result currently ranks first on [SA-23](https://paperswithcode.co/benchmark/sa-23?task=image-segmentation&eval=6194).

The SAM 2 (Hiera-L) result currently ranks second on [MOSE](https://paperswithcode.co/benchmark/mose?task=video-segmentation&eval=6195).

The SAM 2 (Hiera-L) results currently rank third on [DAVIS 2017 val](https://paperswithcode.co/benchmark/davis?task=video-segmentation&eval=6196) and [SA-V test](https://paperswithcode.co/benchmark/sa-v-test?task=video-segmentation&eval=6198).

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/2408.00714/leaderboard-badge-link?eval=6194)
[](https://paperswithcode.co/api/v1/papers/2408.00714/leaderboard-badge-link?eval=6195)
[](https://paperswithcode.co/api/v1/papers/2408.00714/leaderboard-badge-link?eval=6196)
[](https://paperswithcode.co/api/v1/papers/2408.00714/leaderboard-badge-link?eval=6198)
[](https://paperswithcode.co/api/v1/papers/2408.00714/leaderboard-badge-link?eval=6197)

Kind regards,

Niels

Start with the linked paper and its five Papers with Code evaluation pages. Compare the listed scores, model names, benchmark protocols, and openness metadata with the paper and repository, then report or correct any discrepancies; the task is done when all five evaluations are verified.

Written by the indexing model from the issue text.