Hooked CUDA-related dynamic libraries by using automated code generation tools.

Single Shot Multibox Detector by CUDNN6 and C

Pedestrian Target Tracking Based On DeepSORT With YOLOv5

This repository contains the implementation of a real-time human gender detection application using an optimized Darknet Library. For accurate detection of faces closer or farther from the camera, YOLOv3 architecture is used. It was trained on a modified FDDB dataset

Deep neural network. Uses CUDA with cuDNN and cuBLAS_v2 libraries. Provides flexible model building. As an example, classificates cell Images for detecting malaria.

cudnn graph api example