TensorFlow 2.2 implementations for various dynamic neural networks.
- Blockdrop: dynamic inference paths in residual networks
- Dynamic deep neural networks: optimizing accuracy-efficiency trade-offs by selective execution
- Skipnet: learning dynamic routing in convolutional networks
- Multi-scale dense networks for resource efficient image classification (MSDNet)
- Runtime neural pruning
- Stochastic downsampling for cost-adjustable inference and improved regularization in convolutional networks (SDPoint)
- Slimmable Net: Slimmable neural networks; Universally slimmable networks and improved training techniques