Docs: TBD
This repository runs an exported impulse on the Espressif ESP32. See the documentation at Running your impulse locally.
While the script is mainly tested with ESP-EYE, other ESP32-based development boards will work too.
Unzip the deployed C++ library
from your Edge Impulse project and copy only the folders to the root directory of this repository:
example-standalone-inferencing-espressif-esp32/
├─ edge-impulse-sdk
├─ model-parameters
├─ main
├─ tflite-model
├─ .gitignore
├─ CMakeLists.txt
├─ LICENSE
├─ README.md
├─ sdkconfig
├─ sdkconfig.old
└─ partitions.csv
- Initialize ESP IDF:
get_idf
- Compile:
idf.py build
Connect the ESP32 board to your computer.
Run:
idf.py -p /dev/ttyUSB0 flash monitor
Where /dev/ttyUSB0
needs to be changed to actual port where ESP32 is connected on your system.
Use screen, minicom or Serial monitor in Arduino IDE to set up a serial connection over USB. The following UART settings are used: 115200 baud, 8N1.
When switching boards or upgrading to newer version of SDK, the sdkconfig
file in the project folder gets overwritten. Run idf.py menuconfig
to enter configuration menu and make sure that all the relevant performance settings (e.g. Flash SPI speed (80 MHz), CPU Frequency (240 MHz), CONFIG_COMPILER_OPTIMIZATION_PERF=y) are set.