In this sample, you'll see how to call Vertex AI's Gemini Pro in parallel from Workflows. More specifically, you'll gather histories of a list of countries in parallel and return the combined histories in a map.
Make sure you have the right IAM permissions for the default compute service account that Workflows will use:
PROJECT_ID=genai-atamel
PROJECT_NUMBER=$(gcloud projects describe ${PROJECT_ID} --format "value(projectNumber)")
gcloud projects add-iam-policy-binding $PROJECT_ID \
--member serviceAccount:$PROJECT_NUMBER[email protected] \
--role roles/aiplatform.userSee country-histories.yaml or country-histories-connector.yaml for details.
Deploy:
gcloud workflows deploy country-histories-gemini-pro --source=country-histories.yaml
gcloud workflows deploy country-histories-connector-gemini-pro --source=country-histories-connector.yamlRun:
gcloud workflows run country-histories-gemini-pro --data='{"countries":["Argentina", "Brazil", "Cyprus", "Denmark", "England","Finland", "Greece", "Honduras", "Italy", "Japan", "Korea","Latvia", "Morocco", "Nepal", "Oman"]}'
gcloud workflows run country-histories-connector-gemini-pro --data='{"countries":["Argentina", "Brazil", "Cyprus", "Denmark", "England","Finland", "Greece", "Honduras", "Italy", "Japan", "Korea","Latvia", "Morocco", "Nepal", "Oman"]}'You should see an output similar to the following:
