MobileNetV2-YoloV3-Nano: 0.5BFlops 3MB HUAWEI P40: 6ms/img, YoloFace-500k:0.1Bflops 420KB:fire::fire::fire:

A deep learning-powered visual navigation engine to enables autonomous navigation of pocket-size quadrotor - running on PULP

Visualizer for Deep Neural Networks

Cortex M KWS example with Tengine Lite.

Mobilenet v1 trained on Imagenet for STM32 using extended CMSIS-NN with INT-Q quantization support

ffcnn is a cnn neural network inference framework, written in 600 lines C language.

Real time monaural source separation base on fully convolutional neural network operates on Time-frequency domain.

a lightweight js runtime for IOT（一个面向物联网的JS运行时）

OpenCL for Nets - A Deep Learning Framework based on OpenCL, written by C++. Supports popular MLP, RNN(LSTM), CNN(ResNet). Friendly debugger. Transparent data. No library dependencies. 基于OpenCL的深度学习计算框架，C++开发，支持多层感知器，长短时记忆模型，卷积神经网络，残差网络。调试方便，数据透明。无外部依赖。

A python implementation of Patchy-San Convolutional Network for Graph

High Level Synthesis of a trained Convolutional Neural Network for handwritten digit recongnition.

CMix-NN: Mixed Low-Precision CNN Library for Memory-Constrained Edge Devices

RESTful Web Service and C++ compilable version of YOLO written in C and CUDA for object detection.

Face recognition using facenet

使用卷积神经网络在STM32F401C-DISCO上实现人体活动识别

A teeny-tiny CNN that runs on a ESP32-Camera to read digits of a ordinary electricity meter.

Code for ACM MobiCom 2024 paper "FlexNN: Efficient and Adaptive DNN Inference on Memory-Constrained Edge Devices"

Detect dog bark from spectrogram classification using sndfile-tools and darknet

a fast sgemm lib with fix 16 enable on arm 32

A CNN implementation in C.