“Do. Or do not. There is no try.” ― Yoda
Create Your Container Instance
Create a stable, reliable, and reasonably priced cloud container instance suitable for your use.Quick Start
Quickly register and create a Cloud Container Instance by selecting the billing model, host, GPU count, and image.
GPU Selection
Select the right GPU for your tasks. The platform provides proportional CPU and memory resources based on the number of GPUs selected.
Keep Instance Active in Background
Ensure your programs run continuously in the background when using SSH or SSH-based tools by running commands in daemon mode.
Read Data Retention Policies
Understand the data retention rules to avoid data loss, especially during instance shutdowns.
Data Transmission & Storage
We are offering multiple flexible methods for data transfer. It supports various tools and techniques, including JupyterLab, scp commands, and graphical interfaces like FileZilla or XShell, providing users with convenient options for uploading and downloading data.Upload Data
Use various tools like JupyterLab, scp commands, or graphical interfaces for efficient data uploads.
Download Data
Easily retrieve your data using file transfer tools, terminal commands, or intuitive interfaces.
Data Disk
Ideal for intensive tasks, offering flexible and cost-effective storage solutions.
File Storage
Perfect for long-term storage and backup, allowing you to manage data without running an instance.
Enviroment Configuration
The platform’s images come with many frameworks and versions. If they don’t meet your needs, see how to configure other versions.Enviroment Configuration
Customize your environment if the pre-installed frameworks and versions don’t meet your needs.
Image
Save and reuse custom environments by saving images, or share them with other users.
Elastic Deployment
Effortlessly deploy and manage scalable GPU-accelerated applications with intelligent scheduling, rolling updates, and persistent storage — designed for stable, long-running AI workloads.Get Started with Elastic Deployment
Learn how to create ReplicaSets, Jobs, or single containers with flexible resource constraints and automatic host matching.
Why Not Pure Serverless?
Understand why GPUhub chooses elastic containers over traditional serverless for better stability, lower latency, and real-world AI training/inference needs.