A collection of works aiming at deploying NN in embedded systems,inspired by
- DeepMon: Mobile GPU-based Deep Learning Framework for Continuous Vision Applications [MobiSys '17]
- Accelerating Mobile Audio Sensing Algorithms through On-Chip GPU Offloading [MobiSys '17]
- MobiRNN: Efficient Recurrent Neural Network Execution on Mobile GPU [EMDL '17]
- RSTensorFlow: GPU Enabled TensorFlow for Deep Learning on Commodity Android Devices
- Fast and Energy-Efficient CNN Inference on IoT Devices(ucdavis)
- GPU-based Acceleration of Deep Convolutional Neural Networks on Mobile Platforms(ucdavis)
- CNNdroid GPU-Accelerated Execution of Trained Deep Convolutional Neural Networks on Android(ucdavis)
- DeepSense: A GPU-based Deep Convolutional Neural Network Framework on Commodity Mobile Devices
- DeepEye: Resource Efficient Local Execution of Multiple Deep Vision Models using Wearable Commodity Hardware [MobiSys '17]
- DeepX:A Software Accelerator for Low-Power Deep Learning Inference on Mobile Devices(Bell Labs)
- Sparsification and Separation of Deep Learning Layers for Constrained Resource Inference on Wearables(Bell Labs)
- Compression of Deep Convolutional Neural Networks For Fast And Low P ower Mobile Applications(Samsung)
- Learning Structured Sparsity in Deep NeuralNetworks(Pittsburgh)
- Learning both Weights and Connections for Efficient Neural Networks(stanford)
- Faster CNNs With Direct Sparse Convolutions and Guided Pruning(intel)
- Scalpel: Customizing DNN Pruning to the Underlying Hardware Parallelism. (University of Michigan, ARM, ISCA 17)
- Deep Compression:Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding(stanford)
- Ristretto: Hardware-Oriented Approximation of Convolutional Neural Networks(ucdavis)
- SqueezedNet:AlexNet-Level Accuracy With 50x Fewer Parameters and '<'0.5MB Model Size(stanford)
- XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks(Washington)
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications(Google,17)
- ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices(Face++,17)