@bestlibre how did you find that solution? Possibly here

I found the solution based on the curl command in the readme of tensorflow-lite-server and the documentation of requests. The first part of the solution was with a fixed mime type and then I search how to get the mime type from a byte array which lead me to filetype.

Unfortunately on deepstack official this approach raises error: {'success': False, 'error': 'No valid image file found'}

You can run deepstack official with:

docker run -e VISION-DETECTION=True -v localstorage:/datastore -p 5000:5000 -e API-KEY="" --name deepstack deepquestai/deepstack:cpu-x3-beta