# NCNN : High-performance neural network computing

ncnn is a high-performance neural network inference computing framework optimized for embedded platforms. ncnn is designed by Tencent.

# Support

Support most commonly used CNN network:

• Classical CNN: VGG AlexNet GoogleNet Inception ...
• Practical CNN: ResNet DenseNet SENet FPN ...
• Light-weight CNN: SqueezeNet MobileNetV1/V2 ShuffleNetV1/V2 MNasNet ...
• Detection: MTCNN facedetection ...
• Detection: VGG-SSD MobileNet-SSD SqueezeNet-SSD MobileNetV2-SSDLite ...
• Detection: Faster-RCNN R-FCN ...
• Detection: YOLOV2 YOLOV3 MobileNet-YOLOV3 ...
• Segmentation: FCN PSPNet ...

# Features

• Supports convolutional neural networks, supports multiple input and multi-branch structure, can calculate part of the branch
• No third-party library dependencies, does not rely on BLAS / NNPACK or any other computing framework
• Pure C ++ implementation, cross-platform, supports android, ios and so on
• ARM NEON assembly level of careful optimization, calculation speed is extremely high
• Sophisticated memory management and data structure design, very low memory footprint
• Supports multi-core parallel computing acceleration, ARM big.LITTLE cpu scheduling optimization
• Supports GPU acceleration via the next-generation low-overhead vulkan api
• The overall library size is less than 700K, and can be easily reduced to less than 300K
• Extensible model design, supports 8bit quantization and half-precision floating point storage, can import caffe/pytorch/mxnet/onnx models
• Support direct memory zero copy reference load network model
• Can be registered with custom layer implementation and extended

# How to use

Currently neither EdgerOS nor JSRE provides the ncnn API application directly. Users can only use the model operation interfaces provided by EdgerOS and JSRE, including FaceNN and ThingNN ... If you need to use your own neural network model, you can contact us.