Skip to main content

Deployment Quickstart

Deploy your models to production quickly and efficiently.

Deployment Options

Smart Studio offers two main deployment methods to suit different needs and workflows.

Create Deployment

Deploy from the gallery, upload your own, or use a fine-tuned model—all with full configuration control.

  • Browse Model Gallery
  • Upload custom models
  • Advanced configuration
Deploy from Hugging Face

Instantly deploy any public model from Hugging Face Hub using only its model ID. No downloads required.

  • Instant deployment
  • No file uploads
  • Automatic optimization

Quick Deployment Steps

1. Choose Your Deployment Method

Select a custom deployment or deploy from Hugging Face, depending on your needs.

Create Deployment Guide > | Hugging Face Guide >

2. Configure Resources

Select the right GPU, region, and scaling settings to balance performance with cost.

3. Deploy & Monitor

Launch your deployment and monitor performance, usage, and costs through the dashboard.

Best Practices

Resource Selection
  • Match the GPU type to your model's size.
  • Start with a single replica for testing.
  • Choose your deployment region strategically.
Cost Optimization
  • Monitor usage patterns.
  • Set up cost alerts to avoid unexpected charges.

Next Steps

Use Secrets for API Calls

Create API keys to call your deployed models.

Try Model Lab

Test and experiment with your deployed models in our interactive playground.

Post-Training

Post-train your models with custom data for better performance.