Hello, I got my openMV N6 device recently and I am experiencing the below issue with it while loading the ML models. Below are the complete details of the issue.
Device used: OpenMV N6 (STM32N657X0, Cortex-M55 + Neural-ART NPU)
IDE Version: OpenMV IDE 4.8.9 on Windows 11
Current Firmware: v4.8.1, MicroPython v1.26.0-77, compiled 2025-12-22
Issue Background: The original factory firmware worked perfectly with ML models. I successfully ran 3 combined ML object models using ROM models. The problem started after I manually installed the development build (v1.28.0, compiled 2026-04-16) via Tools → Install Latest Development Release, while working on an audio project. After that update, whenever any script causes a hard fault, the IDE detects DFU mode and automatically reflashes back to its bundled v1.26.0-77 (v4.8.1).
Problem: After the IDE then automatically reflashed its bundled firmware (v4.8.1, MicroPython v1.26.0-77, compiled 2025-12-22). the firmware cannot load any ML model at all, so every attempt to run an ML script hard faults, triggering another reflash of the same broken firmware.
What triggers the hard fault: Even the most minimal possible ML script on a completely fresh boot like below triggers the fault and DFU mode:
import ml
model = ml.Model(“/rom/yolov8n_192.tflite”)
print(model)
What works fine:
-
Basic camera (CSI) scripts work perfectly
-
helloworld with
import csiruns without issues -
ROM contents are intact - all 17 model files present including
yolov8n_192.tflite,hand_landmarks_full_224.tflite,movenet_singlepose_192.tflite, etc. -
All non-ML scripts work fine
Can you please guide me on how to fix the issue I am experiencing on my N6 device? Is there a way to restore the original factory firmware that shipped with the N6?
Thanks,
Swathi


