MNN
Run converted neural networks in local mobile and desktop applications using the MNN runtime.
Published by Alibaba
How to use it
Convert a compatible model to MNN format, then load it from a native application with the runtime for the target operating system. Include preprocessing and postprocessing in your equivalence test.
Choose a CPU or supported GPU backend for each build. Verify operator coverage and latency on the device you will ship because GPU support depends on the platform and compiled backend.
- Hardware
- CPU, OpenCL, Vulkan, Apple Metal
- Model formats
- MNN
- Operating systems
- Android, iOS, Linux, Windows, macOS
- Access
- Runs locally. Supply a compatible MNN model and target-specific runtime build.