Various MLTK commands such as profile and classify_audio support executing on the desktop (e.g. Windows/Linux) or on supported embedded platforms. Additionally, C++ applications may be developed for the desktop as well as supported embedded platforms.
The following embedded platforms are currently supported:
- Name: EFR32xG24 Dev Kit
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd2601)
This platform features the MVP machine learning hardware accelerator.
To build a C++ application with MVP hardware acceleration using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(TFLITE_MICRO_ACCELERATOR mvp)
The platform also supports the following commands when using --device command line option:
- Name: EFM32 Giant Gecko S1, GG11 Starter Kit
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd2204)
The platform also supports the following commands when using --device command line option:
- Name: Thunderboard Sense 2
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd4166)
The platform also supports the following commands when using --device command line option:
- Name: EFR32xG24 Pro Kit +10 dBm
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd4186)
This platform features the MVP machine learning hardware accelerator.
To build a C++ application with MVP hardware acceleration using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(TFLITE_MICRO_ACCELERATOR mvp)
The platform also supports the following commands when using --device command line option:
- Name: EFR32xG28 2.4 GHz BLE and +20 dBm Radio Board
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd4401)
This platform features the MVP machine learning hardware accelerator.
To build a C++ application with MVP hardware acceleration using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(TFLITE_MICRO_ACCELERATOR mvp)
The platform also supports the following commands when using --device command line option:
- Name: EFR32xG28 Explorer Kit
- Product Link
To build a C++ application for this platform using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(MLTK_PLATFORM_NAME brd2705)
This platform features the MVP machine learning hardware accelerator.
To build a C++ application with MVP hardware acceleration using VSCode or Command-Line, create/modify <mltk repo root>/user_options.cmake:
mltk_set(TFLITE_MICRO_ACCELERATOR mvp)
The platform also supports the following commands when using --device command line option: