Add NNCLR pre-trained models

[Rishit-dagli]

NNCLR is a self-supervised learning approach as proposed in the paper "With a Little Help from My Friends: Nearest-Neighbor Contrastive Learning of Visual Representations" [1], by Google Research and DeepMind pre-trained on STL-10.

The training code was contributed by me on Keras Examples and could be found here.

References

[1] Dwibedi, Debidatta, et al. “With a Little Help from My Friends: Nearest-Neighbor Contrastive Learning of Visual Representations.” ArXiv:2104.14548 [Cs], Apr. 2021. arXiv.org, http://arxiv.org/abs/2104.14548

---

Any pull request you open is subject to the TensorFlow Hub Terms of Service at www.tfhub.dev/terms and Google's Privacy Policy at https://www.google.com/policies/privacy.

We check modified Markdown files for validity using this GitHub Workflow. You can execute the same checks locally by passing the file paths of modified Markdown files (relative to the assets/docs directory) to validator.py e.g. python3 ./tools/validator.py google/google.md google/models/albert_base/1.md ....

[github-actions]

By opening this pull request, you are agreeing to the TensorFlow Hub Terms of Service at www.tfhub.dev/terms and Google's Privacy Policy at https://www.google.com/policies/privacy, and your pull request is being reviewed.

[WGierke]

Thanks @Rishit-dagli for providing this model! Would you mind addressing the one comment I left?

[Rishit-dagli]

Thanks @WGierke for the review, added those.

[github-actions]

Thank you for your contribution. Your pull request has been accepted according to the TensorFlow Hub Terms of Service at www.tfhub.dev/terms and Google's Privacy Policy at https://www.google.com/policies/privacy. Your model should appear on tfhub.dev within a day.