mit-han-lab/tinyengine
A memory-efficient neural network library for running and training deep learning models on microcontrollers.
Not currently ranked — collecting fresh signals.
star history
TinyEngine is part of the MCUNet framework for tiny deep learning on IoT devices. It provides an inference runtime optimized for memory-constrained microcontrollers, supporting quantized neural networks and patch-based inference to fit within tight memory budgets. The library also enables on-device training under 256KB memory, with demos for person detection and face mask detection on OpenMV cameras.
Frequently asked
- What is mit-han-lab/tinyengine?
- A memory-efficient neural network library for running and training deep learning models on microcontrollers.
- Is tinyengine open source?
- Yes — mit-han-lab/tinyengine is open source, released under the MIT license.
- What language is tinyengine written in?
- mit-han-lab/tinyengine is primarily written in C.
- How popular is tinyengine?
- mit-han-lab/tinyengine has 951 stars on GitHub.
- Where can I find tinyengine?
- mit-han-lab/tinyengine is on GitHub at https://github.com/mit-han-lab/tinyengine.