ModusToolbox™ for Machine Learning
ModusToolbox™ for Machine learning covers three primary areas for product makers:
- Bring-your-own Model
- Train-your-own Model
- Buy-your-own Model
ModusToolbox™ Machine Learning (ML) enables you to rapidly evaluate and deploy Machine Learning models on Infineon MCUs. ModusToolbox™ ML is designed to work seamlessly with the ModusToolbox™ ecosystem and can be added into existing projects to enable inferencing on low-power edge devices.
Easy Integration and Flexible Tools
ModusToolbox™ ML tools enable you to:
- Import models from popular training frameworks such as TensorFlow™
- Optimize the model for embedded platforms to reduce the size and complexity
- Validate the performance of the optimized model by checking performance against test data
- Generate optimized model code and libraries which are integrated with the ModusToolbox™ flow
Optimized for Embedded Performance
ModusToolbox™ ML includes an embedded inference engine which supports optimized implementations of most popular Neural Network operators such as 1D/2D convolutions, a variety of activation functions as well as support for more complex operators for RNN networks such as GRU.
ModusToolbox™ ML adds configurators, tools, code examples and supporting libraries to help quickly get started with Machine Learning Model deployment on Infineon MCU’s.
ModusToobox™ ML provides a set of featuresthat make the deployment and validation of Deep Learning models on the PSoC™ 6 platform seamless. These features include:
- TensorFlow Lite for Microcontrollers including both Interpreter and interpreter-less inferencing
- Supports .tflite and .H5 model formats
- Supports the following characteristics of NNs:
- Core NN Kernels: MLP, GRU, Conv1d, Conv2d, LSTM
- Support NN Kernels: flatten, dropout, reshape, input layer
- Activations: relu, softmax, sigmoid, linear, tanhInput Data Quantization Level:
- 32-bit float
- 16/8-bit integer
- NN Weights Quantization Level:
- 32-bit float
- 16/8-bit integer
- Regression Data Evaluation
- Cycle and memory estimation
- PC based inference engine
- Target device-based inference engine (optimized)
ModusToolbox™ for Machine Learning also incorporates partners that combine the best of Infineon Software and Partner Software through our ModusToolbox™ and Friends Ecosystem. The partners provide simple to use AutoML training platforms allowing developers to focus on their data and use cases instead of the heavy infrastructure and learning curve traditional Machine Learning training pipelines incur.
Infineon works with the following partners to combined ModusToolbox and their AutoML Platform.
Edge Impulse are a cloud-based Machine Learning training platform that is deployable to TinyML devices. With Edge Impulse’s partner ecosystem and expansive training content, deploying machine learning models to embedded microcontrollers is made easier. To get more information on Edge Impulse, go here. To get started with the Infineon and Edge Impulse integration, please go here.
ModusToolbox™ for Machine Learning also brings the concept of Buy-your-own Model which is targeted towards product makers who do not need customization, but just need to deploy something that works.
Infineon collaborates Cyberon to provide an offline keyword spotting implementation that can be easily deployed to Infineon Microcontrollers. To get started, please go here.