Seeed Grove Vision AI Module V2 gets stuck with anomaly detection model,

Question/Issue:
[Hi,
I’ve created a anomaly detection model for the Seeed Grove Vision AI Module V2 board. When I try the model with the webCam or the smartphone, it works pretty fine, but when I use on the board it self, it takes the photo and gets stuck with no result or error.
Base on the training output performance results, it should require 194K of RAM and 117K of flash.
I’ve tried with different resolutions, and datasets of different quality, but the result is the same.
Thank you for your time and help!]

Project ID:
[713737]

Context/Use case:
[Created a model with ]

Steps Taken:

  1. [Data acquisition with multiple devices (Around 180 images) (Tried labeling and no)]
  2. [Create impulse with image size of 96*96 (fit longest axis) and FOMO-AD]
  3. [Train success]
  4. [Deploy using Seed Grove Vision AI Module V2 profile with Quantized.]
  5. [Flash using the following command: python3 xmodem/xmodem_send.py --port=/dev/ttyACM0 --baudrate=921600 --protocol=xmodem --file=firmware.img --model=“model_vela.tflite 0x200000 0x00000”]
  6. [It doesn’t reboot, when you push y. Manual reboot pushing on board button.]
  7. [Run Impulse with edge-impulse-run-impulse command]

Expected Outcome:
[It should return a traces with the result of the anomaly.]

Actual Outcome:

Reproducibility:

  • [X] Always
  • [ ] Sometimes
  • [ ] Rarely

Environment:

  • Platform: [Seeed Grove Vision AI Module V2.]
  • Build Environment Details: [Seeed Grove Vision AI Module V2]
  • OS Version: [Ubuntu 22.04]
  • Edge Impulse Version (Firmware): [e.g., 1.72.17]
  • To find out Edge Impulse Version:
  • if you have pre-compiled firmware: run edge-impulse-run-impulse --raw and type AT+INFO. Look for Edge Impulse version in the output.
  • if you have a library deployment: inside the unarchived deployment, open model-parameters/model_metadata.h and look for EI_STUDIO_VERSION_MAJOR, EI_STUDIO_VERSION_MINOR, EI_STUDIO_VERSION_PATCH
  • Edge Impulse CLI Version: [e.g., 1.5.0]
  • Project Version: [e.g., 1.0.0]
  • Custom Blocks / Impulse Configuration: [Describe custom blocks used or impulse configuration]
    Logs/Attachments:
    [Edge Impulse impulse runner v1.32.1
    [SER] Connecting to /dev/ttyACM0
    [SER] Serial is connected, trying to read config…
    [SER] Retrieved configuration
    [SER] Device is running AT command version 1.8.0

[SER] To get a live feed of the camera and live classification in your browser, run with --debug

[SER] Started inferencing, press CTRL+C to stop…
LSE
cis_OV5647_init
mclk DIV3, xshutdown_pin=82
Set PA1(AON_GPIO1) to High
hx_drv_cis_set_slaveID(0x36)
OV5647 Init Stream by app
Allocating for subs 1
WD1[34045ac0], WD2_J[34045ac0], WD3_RAW[3404a5c8], JPAuto[34043e40]
MIPI TX CLK: 96M
MIPI CSI Init Enable
MIPI TX CLK: 96M
MIPI BITRATE 1LANE: 440M
MIPI DATA LANE: 2
MIPI PIXEL DEPTH: 10
MIPI LINE LENGTH: 640
MIPI FRAME LENGTH: 480
MIPI CONTINUOUSOUT: 1
MIPI DESKEW: 0
t_input: 7387ns
t_output: 6666ns
t_preload: 385ns
MIPI RX FIFO FILL: 10
MIPI TX FIFO FILL: 0
RESET MIPI CSI RX/TX
VMUTE: 0x00000000
0x53061000: 0x0606070D
0x53061004: 0x2B07201C
0x53061008: 0x00000505
0x5306100C: 0x00005500
0x53061010: 0x00000000
sensor_type: 15
OV5647 on by app done
g_cursensorstream: 0
Inferencing settings:
Image resolution: 96x96
Frame size: 9216
No. of classes: 0
Starting inferencing in 1 seconds…
cis_OV5647_init
mclk DIV3, xshutdown_pin=82
Set PA1(AON_GPIO1) to High
hx_drv_cis_set_slaveID(0x36)
OV5647 Init Stream by app
Allocating for subs 1
WD1[340829d0], WD2_J[340829d0], WD3_RAW[340874d8], JPAuto[34043ee0]
MIPI TX CLK: 96M
MIPI CSI Init Enable
MIPI TX CLK: 96M
MIPI BITRATE 1LANE: 440M
MIPI DATA LANE: 2
MIPI PIXEL DEPTH: 10
MIPI LINE LENGTH: 640
MIPI FRAME LENGTH: 480
MIPI CONTINUOUSOUT: 1
MIPI DESKEW: 0
t_input: 7387ns
t_output: 6666ns
t_preload: 385ns
MIPI RX FIFO FILL: 10
MIPI TX FIFO FILL: 0
RESET MIPI CSI RX/TX
VMUTE: 0x00000000
0x53061000: 0x0606070D
0x53061004: 0x2B07201C
0x53061008: 0x00000505
0x5306100C: 0x00005500
0x53061010: 0x00000000
sensor_type: 15
OV5647 on by app done
g_cursensorstream: 0
OV5647 on by app done
g_cursensorstream: 0
Taking photo…
]

[> AT+INFO
AT+INFO


  • Edge Impulse firmware *

Firmware build date : Jun 30 2025
Firmware build time : 08:16:13
ML model author : Juan
ML model name : anomalyTest
ML model ID : 732504
Model deploy version : 1
Edge Impulse version : v1.72.17
Used sensor : camera

]

Additional Information:
[This issue only happens with anomaly detection models, if I try a object detection model, it works fine]