from django.conf import settings from roboflow import Roboflow import os rf = Roboflow(api_key=os.environ["API_KEY_ROBO"]) model = rf.workspace().project("plankton-vhsho").version(1).model def predict_image(image): results = model.predict(image.image.path) results.save( f"{settings.MEDIA_ROOT}/{image.image.name.split('.')[0]}_predicted.{image.image.name.split('.')[-1]}" ) return results.json()