A fork of Intrinsic image popularity Assessment
This is a PyTorch implementation of the paper Intrinsic Image Popularity Assessment.
Run python src/rateImage.py -path <path-to-directory> -e <image-ext>
to evaluate the intrinsic image popularity of your photos on Instagram.
Follow the cloudbuild.yaml file
export SERVICE_URL_TAGS=$(gcloud run services describe iipa --format "value(status.traffic.tag)")
export SERVICE_URL=$(gcloud run services describe iipa --format "value(status.url)")
export SERVICE_ACCOUNT=$(gcloud run services describe iipa --format "value(spec.template.spec.serviceAccountName)")
export DJANGO_SUPERUSER_PASSWORD=$(gcloud secrets versions access latest --secret=superuser_password)
gcloud builds submit --config cloudbuild.yaml --substitutions _SERVICE_URL_TAGS=$SERVICE_URL_TAGS,_CLOUDRUN_SERVICE_URL=$SERVICE_URL,TAG_NAME=v0.10.5
gcloud run deploy iipa --image gcr.io/iipa-32fdd/iipa:latest --service-account $SERVICE_ACCOUNT --set-env-vars DEBUG=False,GCP_DEV=True,SERVICE_URL_TAGS=$SERVICE_URL_TAGS,CLOUDRUN_SERVICE_URL=$SERVICE_URL
- make sure Firebase Hosting is rewriting reqs correctly