- GPU Instance: A dedicated virtual machine with full control over GPU hardware, ideal for long-running, predictable workloads.
- Serverless GPUs: A fully managed, on-demand GPU service that dynamically scales and charges based on actual usage, perfect for burstable and sporadic tasks.
Create your first GPU instance
Here’s a guide on how to create, manage and use your first GPU instance.
Set up a Serverless Endpoint from Scratch
To learn about the Serverless GPUs, follow this guide.