Resource Management Error: Building ResNet8 TFLite model on Nucleo-N657X0-Q (X-CUBE-AI)

[Sradha]

Hi,

I am a beginner working with the STM32 Nucleo-N657X0-Q and trying to deploy a pre-trained model from the ST Model Zoo.

Environment:

• Board: Nucleo-N657X0-Q
• Tools: STM32CubeMX (latest), STM32CubeIDE (latest), X-CUBE-AI (latest)
• Model: resnet8_32_tfs_int8.tflite
• Workflow: Followed the STM32CubeIDE 2.0.0 workflow tutorial from the ST Community. (STM32CubeIDE 2.0.0 workflow tutorial - STMicroelectronics Community)

Configurations:

• Heap Size: 0x4000
• Stack Size: 0x2000
• Toolchain/IDE:STM32CubeIDE
• Used default X-CUBE-AI settings for code generation. 

The Issue:
After generating the C code in CubeMX and attempting to build the project in STM32CubeIDE, the build fails with the following error

Steps Taken:

1. Imported the TFLite model into the X-CUBE-AI simplified GUI.
2. Verified the model successfully (Analysis report shows it should fit).
3. Generated the code and opened the project in STM32CubeIDE.
4. Attempted to build. 

Could someone guide me on whether my Heap/Stack allocations are insufficient for the model, or if there are additional linker script adjustments/additional configuration change needed for the N6 series?

Attached are the screenshots of my Cube MX configuration and the console error output. Any help would be greatly appreciated

[Julian E.]

Hi @Sradha,

 

I would suggest to use the deployment operation model of the model zoo services for that:

https://github.com/STMicroelectronics/stm32ai-modelzoo-services

 

Also X Cube AI is deprecated, we advise you to now use STM32Cube AI Studio:

https://community.st.com/t5/edge-ai/meet-stm32cube-ai-studio-the-new-go-to-tool-for-embedded-ai/td-p/880369

 

Note that for the Nucleo N6, a bug is identified and will be fixed in next release (which should come in the month).

 

Have a good day,

Julian