tf-DynamincNNs

TensorFlow 2.2 implementations for various dynamic neural networks.

Model List

1. Blockdrop: dynamic inference paths in residual networks
2. Dynamic deep neural networks: optimizing accuracy-efficiency trade-offs by selective execution
3. Skipnet: learning dynamic routing in convolutional networks
4. Multi-scale dense networks for resource efficient image classification (MSDNet)
5. Runtime neural pruning
6. Stochastic downsampling for cost-adjustable inference and improved regularization in convolutional networks (SDPoint)
7. Slimmable Net: Slimmable neural networks; Universally slimmable networks and improved training techniques