Dockerfile for Raspberry Pi3/4

[koeswt]

Hi,

when exporting a Custom Vission Dockerfile to RPi4 (i.e. the Dockerfile you get for Raspberry Pi 3) some image processing libraries are missing (e.g. libtiff5.so) and there is versioning trouble with numpy. Here is what worked for me on an RPi4 with 32 Bit Bullseye:

FROM arm32v7/python:3.9-slim-bullseye RUN apt update && apt install -y libjpeg62-turbo libopenjp2-7 libtiff5 libatlas-base-dev libgl1-mesa-glx RUN pip install --no-cache-dir "flask<3" "pillow<11" "numpy==1.21.2" tflite-runtime~=2.13.0 --extra-index-url=https://www.piwheels.org/simple

I suggest you add an export option for RPi4. I suppose one has the same trouble with export to RPi3.

Best regards

KOE

Document Details

⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.

• ID: 8198ccf7-c499-c84b-6d54-e48a8fc4a150
• Version Independent ID: aae607bf-72be-59dd-32bc-be33039de820
• Content: Quickstart: Build an image classification model with the Custom Vision portal - Azure AI services
• Content Source: articles/ai-services/custom-vision-service/getting-started-build-a-classifier.md
• Service: azure-ai-custom-vision
• GitHub Login: @PatrickFarley
• Microsoft Alias: pafarley

[TPavanBalaji]

@koeswt Thanks for your feedback! We will investigate and update as appropriate.

[AjayBathini-MSFT]

@koeswt Thanks for your feedback! I've assigned this issue to the author who will investigate and update as appropriate.

[PatrickFarley]

Hi @koeswt , thank you for your feedback, I'll add this ask to our backlog

please-close