nndeploy
nndeploy is an easy-to-use and high-performance AI deployment framework designed to deploy AI algorithms on edge devices, including desktop platforms (Windows, macOS), mobile devices (Android, iOS), edge computing hardware (NVIDIA Jetson, Ascend310B, RK series), and standalone servers (RTX series, T4, Ascend310P). It features visual workflow design with drag-and-drop nodes for building AI pipelines with adjustable parameters, supports custom nodes in Python and C++/CUDA, and enables one-click export of workflows as JSON for C++/Python API integration. For large models exceeding 10B parameters such as large language models and AIGC generation models, nndeploy serves as a visual workflow tool. Performance features include serial, pipeline parallel, and task parallel execution modes, zero-copy memory optimization, memory pooling, and built-in C++/CUDA/Ascend C/SIMD optimized nodes. It integrates 13 mainstream inference frameworks including ONNXRuntime, TensorRT, OpenVINO, MNN, TNN, and ncnn, allowing flexible in