Serverless GPU

Float16 Serverless GPU lets you run AI workloads on powerful GPUs, paying only for the time you use. No infrastructure management required.

How It Works

1. Create a project - Set up a project in the dashboard or via CLI
2. Write your code - Use Python to run your AI workloads
3. Run tasks - Execute your code on GPU instances
4. Pay per second - Only pay for actual compute time

Getting Started

Install Float16 CLI

The easiest way to get started is with the VS Code extension:

1. Open VS Code
2. Install the Float16 CLI Tools extension
3. The extension handles authentication and project setup automatically

Alternatively, install via command line:

Create a Project

1. Navigate to Serverless GPU > Projects
2. Click Create Project
3. Select a GPU instance type
4. Enter a project name (optional)
5. Click Create Project

Pricing

View current pricing at Serverless GPU > Pricing.

Blueprint Templates

Get started quickly with pre-configured templates at Serverless GPU > Blueprint:

Managing Projects

Projects

View and manage your serverless GPU projects at Serverless GPU > Projects.

Tasks

Track running and completed tasks at Serverless GPU > Tasks:

• View task history
• Monitor execution status
• Access task outputs

Tokens

Manage API tokens for CLI authentication at Serverless GPU > Tokens.

Storage Logs

Track storage operations at Serverless GPU > Storage Logs:

• Upload/download history
• Operation statistics
• 7-day activity charts

Features

Streaming Responses

Get real-time output as your code executes in manual mode.

VS Code Integration

Run and manage your Serverless GPU projects directly from VS Code with the Float16 CLI Tools extension.

Next Steps

• Blueprints - Learn more about templates
• GPU Platform - For dedicated GPU instances