Failed to allocate TFLite Arena

[jdbendell]

Hi,

Question/Issue:
I’m getting the failed to allocate arena error.

I’ve tried the tips in these posts, and they don’t seem to fix it. What should I try doing?

“Failed to allocate TFLite arena (error code 1)” when using FOMO for centroid detection

I am experimenting with using FOMO for centroid detection with an Arduino Nano 33 BLE Sense (Cortex-M4F 64MHz) and an OV7670 camera. I’m using the Arduino IDE 1.8.19 on a laptop with Ubuntu 18.04. I have made the following settings in the IDE: Tools->Board->Board Manager->Arduino Mbed OS Nano Boards (3.4.1) Tools->Board->Arduino Mbed OS Nano Boards->Arduino Nano 33 BLE Tools->Port->/dev/ttyACM0 Tools->Programmer->ARM CMSIS-DAP compatible Arduino AVR Boards (1.8.6) My training/test data is 1…

OV767X support with Arduino Nano

I just tried the new example “nano_ble_sense_camera” that was generated using TL (MobileNetV1 0.2 (final layer: 16 neurons, 0.1 dropout). Doing the inference, the photo is captured only once and the code is frozen on Arduino. [image] Sometimes, after reset, an error occurred, regarding TFLite arena allocation [image] I run the same model on a modified static_buffer example code and it worked well. [image] Here is possible to see one of the images: [image] The code that I used was dev…

Project ID:
163740

Context/Use case:
I’m attempting to do object classification on an Arduino Nano 33 BLE Sense with the OV7675 camera. I’m using MobileNet .35 and 78x78 images. I’m downloading the quantized model which says 174 K RAM and 75K flash. This should fit on the Arduino right?

Also, I’m a student at UPenn, and this is for a class final project due tomorrow. Still haven’t learned not to procrastinate.

Any help is much appreciated!

[jdbendell]

[jdbendell]

Figured it out. Wahoo. Shrunk the image size down to 50 by 50.

[louis]

@jdbendell

Also, I’m a student at UPenn, and this is for a class final project due tomorrow. Still haven’t learned not to procrastinate.

Anyway, glad that you figured it out.
For the 174K RAM, it is an estimation and keep in mind that the board is also using some RAM for other purposes. Here’s a tutorial that might interest you: Tracking memory usage with Mbed OS | Mbed

Best,

Louis