I am using FOMO to detect humans, but I keep getting the following error: ERR: failed to allocate tensor arena
Failed to allocate TFLite arena (error code 1)
Failed to run impulse (-6)
I looked into it, and found that I need to reduce the model peak RAM usage, so now it is 141.3k, I am using Arduino Nano 33 BLE Sense, which has a 250k RAM, so it should be enough but I keep getting the same error.
any suggestions are welcome, it’s URGENT.
What is your Input Image Block set to?
You’ll probably need to goto an Impulse design that has an Input Image Block set to 96x96 that in turn will reduce the RAM usage you see on the Deployment tab in the EI Studio.
The estimates you see in the EI Studio are estimates for your Model’s usage and performance. Your INO file will use RAM and Flash as well.
Ok so I tried to reduce the input image size, and the one that worked is 66 x 66 which is very low, and is negatively affecting my F1 score and accuracy. Any way I can get lower RAM usage without having to reduce the image size that low?
@Mariam_Shahoud as @MMarcial mentioned, one thing to understand is that the RAM consumption in the studio is for the FOMO Model itself. This does not include the firmware RAM consumption, which is the total you will need to sum both.
I would recommend using something more suitable for FOMO such as the Arduino Nicla vision