hardware code developed based on platformIO (a vscode extansion). Just open the hardware folder as project and adjust some code to satisfy you device and build/upload.
spend of memory:
Advanced Memory Usage is available via "PlatformIO Home > Project Inspect" RAM: [== ] 23.8% (used 77852 bytes from 327680 bytes) Flash: [========= ] 87.5% (used 1147233 bytes from 1310720 bytes)
in order to use the screen to show the progress, you need to modify the user_setup.h in the hardware folder and copy it into the correct position
the model was trained on my computer cpu based on tensorflow-cpu. and the data was processed and upload to the server. Also build a server to send data to my ESP32 to get data.
the python dependency mainly use: tensorflow-cpu==2.11.0 numpy tqdm