Setting up backend infrastructure for model deployment using vertex ai - Githubissues

opencampus-sh / sign-language-translator

German Sign Language Translator

GNU General Public License v3.0

1 stars 0 forks source link

Setting up backend infrastructure for model deployment using vertex ai #18

Open steffen74 opened 5 days ago

steffen74 commented 5 days ago

The setup for the model deployment should be in accordance with the following overall deployment scenarios:

Development

Uses CPU-only resources
Single replica deployment
Mock model available for offline development
Minimal compute resources for cost optimization

Staging

GPU-enabled endpoints
Automated deployment via GitHub Actions
Integration testing before promotion
Scaled-down production configuration

Production

Full GPU resources
Multiple replicas for high availability
Automated canary deployments
Full monitoring and alerting