Facebook AI open-sources Opacus, a new high-speed library for training PyTorch models with differential privacy (DP)
Opacus is a library that enables training PyTorch models with differential privacy. It supports training with minimal code changes required on the client, has little impact on training performance and allows the client to online track the privacy budget expended at any given moment.
https://github.com/pytorch/opacus?fbclid=IwAR3Vsr9jSpHFDq0HMpPerxnMeHZ0CdqzoGDfasbXGX61TntAhiTywQQ8UQw

Deep learning is a field with intense computational requirements, and your choice of GPU will fundamentally determine your deep learning experience. But what features are important if you want to buy a new GPU? GPU RAM, cores, tensor cores? How to make a cost-efficient choice? 
https://timdettmers.com/2020/09/07/which-gpu-for-deep-learning/

Recently PyTorch has gained a lot of popularity because of its ease of usage and learning. Andrej Karpathy, Senior Director of AI at Tesla, said the following in his tweet.

OpenCV 4.5.2 and 3.4.14 have been released!

https://opencv.org/opencv-4-5-2/

Release highlights for 4.5.2

1. Added support for dynamically loaded parallel_for backends

2. Added IntelligentScissors algorithm implementation

3. Improvements in dnn module: 3(a) supported several new layers: Mish ONNX subgraph, NormalizeL2 (ONNX), LeakyReLU (TensorFlow) and others 3(b) supported OpenVINO 2021.3 release

4. G-API module got improvements in inference and media processing areas

5. Improved hardware-accelerated video decoding and encoding 6. Added Android NDK camera support

7. Added WeChat QRCode module to the opencv_contrib

“If 80 percent of our work is data preparation, then ensuring data quality is the important work of a machine learning team.”~ Andrew Ng

New computer vision system achieving state of the art results in image segmentation, without needing any labeled training data. This new model was trained on random, unlabeled data, but quickly achieved state-of-the-art results

https://github.com/facebookresearch/dino

There are some issues with facial recognition. It doesn't work so well with black people.
https://m.youtube.com/watch?v=nGStQVeCYuw&feature=youtu.be

Deep learning for Computer Vision

The techniques covered in this
book are meant to be much more advanced than the Starter Bundle – this is where you’ll start to
separate yourself from a deep learning novice and transform into a true deep learning practitioner. To start your transformation to deep learning expert, just flip the page of this wonderful book by Dr. Adrian Roserbrok

Big announcement today by #PyImageSearch 📢

Everyone, everywhere gets FREE unlimited access to PyImageSearch University until May 24th 💥🚀😎

18 courses · 91 lessons · 29h 21m video

Join here: https://bit.ly/3w30adj 💪

Please share this post with anyone interested in learning Computer Vision, Deep Learning, and OpenCV.

TensorFlow 2.5.0 has been released with wide variety of features and some bug fixes and improvements. For More details check: https://github.com/tensorflow/tensorflow/releases

💃 Dance, jump, and twirl! 🤸

TensorFlow just released MoveNet, a new lightning-fast pose model in browser that's #MadeWithTFJS and designed to track fast motions and tough poses.

Learn more at #GoogleIO and try it yourself → https://t.co/QpfnVL0YYI https://t.co/zZpvc43aEP

PyTorchVideo: A deep learning library for video understanding
https://ai.facebook.com/blog/pytorchvideo-a-deep-learning-library-for-video-understanding/

Excited to announce PyTorch Enterprise for Azure, and to partner with Facebook on the new PyTorch Enterprise Support Program!

https://venturebeat-com.cdn.ampproject.org/c/s/venturebeat.com/2021/05/25/facebook-and-microsoft-launch-pytorch-enterprise-support-program/amp/