Google MorphNet makes your neural net faster and smaller
Facebook segments objects in image or video with no supervision
Removing haze in images
Unsupervised shape completion of 3D data
Get summary video from a text search
Multiscale Vision Transformers Outperform Prior Work
Survey of deep learning to enhance dim images
Shapash library for making ML models understandable
Kornia Computer Vision Library for PyTorch
Ensemble deep learning: A review
Transformers for action recognition 40 times faster by focus attention in time
NVIDIA, Stanford, Microsoft Efficient Trillion-Parameter Model Training on GPU Clusters
15 graphs show state of AI in 2021
Transformer for face recognition comparable to CNN
Microsoft DeepSpeed lets 1 GPU train 40 billion parameter neural net
Fabric with display and touch input available next year
Vision transformer for object detection beats state-of-art ResNet-50
CPU beats GPU by 15x for neural net training with Rice Univ method
Fourier transformers for super-resolution images and CT scans
Top 10 AI and Machine Learning Papers in 2020