gemini-test

Project description

In this project, we test gemini AI with multimodal prompts. The supported model is gemini-pro-vision.

Configure Google Cloud

1. Complete the following configuration according to Google Cloud Doc.

• Select a proper Google Cloud project
• Enable the Vertex AI API
• Install the Google Cloud CLI
• Update and install gcloud components

2. Create a service account and generate a service account key according to this document: https://cloud.google.com/vertex-ai/docs/start/client-libraries#before_you_begin

3. Install Google CLoud Client Library

npm install @google-cloud/aiplatform

4. Configure your own google cloud service account credential in backend

process.env.GOOGLE_APPLICATION_CREDENTIALS = 'Your Own Service Account Credential Path';

Start the frontend

npm run serve

Start the backend

node index.js

References

1. Send multimodal prompt requests with images