In deployment option the ST BL475E IOTA1 Board is showing not supported
Hi @swathi.2003,
My guess is that you configured a project in Studio that does not run on the ST BL475E IOTA1 board (so the option is grayed out on the deployment page). If you provide us with your project ID number, we can take a look for you.
1 Like
Hi @swathi.2003,
If you search for “ST” in the deployment options, you can see that the ST BL475E IOTA1 (also known as the “ST IoT Discovery Kit”) is grayed out. The pre-compiled firmware for that board does not contain code for a camera, so vision examples will not work.
You have a few options from here:
- Modify the firmware to support a camera or images given in the link in the grayed out board option: GitHub - edgeimpulse/firmware-st-b-l475e-iot01a: Edge Impulse firmware for the ST B-L475E-IOT01A development board
- Deploy as a Cube.MX CMSIS-PACK and import it into your STMicro IDE as a library: https://docs.edgeimpulse.com/docs/run-inference/using-cubeai
- Deploy as a C++ library, import it into your IDE, and link to it. Here is an example of how I did that with a keyword spotting project: https://github.com/ShawnHymel/ei-keyword-spotting/tree/master/embedded-demos/stm32cubeide/nucleo-l476-keyword-spotting
1 Like
Hi @swathi.2003,
“Failed to allocate tensor arena” means that you have run out of RAM on your microcontroller while initializing the ML model. You have a few options:
- Reduce the input size (resolution) of your image
- Try a different/smaller model
- Use a different microcontroller
See here for a similar discussion (different board): Failed to allocate tensor arena on Nano.ble33_sense