FasterDNN

Deep Neural Networks(include DNN/CNN/RNN) run on mobile devices

Paper	Compression	Pruning	Software	Hardware	DNNs
Learning both Weights and Connections for Efficient Neural Networks	√	〤	×		DNN/CNN
Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding	√	√	×		DNN/CNN
EIE: Efficient Inference Engine on Compressed Deep Neural Network	√		√	√	DNN
Improving the speed of neural networks on CPUs	√		√		DNN
RESHAPING DEEP NEURAL NETWORK FOR FAST DECODING BY NODE-PRUNING		√			DNN
ON THE COMPRESSION OF RECURRENT NEURAL NETWORKS WITH AN APPLICATION TO LVCSR ACOUSTIC MODELING FOR EMBEDDED SPEECH RECOGNITION	√				RNN
Optimizing Performance of Recurrent Neural Networks on GPUs			√	√	RNN
Restructuring of Deep Neural Network Acoustic Models with Singular Value Decomposition	√				DNN
SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size	√	√			CNN

Automatic Speech Recognition