Compute Power Usage
Platform Features
The platform provides mainstream AI inference application models and deep learning frameworks, supporting one-click deployment that's ready to use out of the box. No manual environment configuration is required, significantly reducing deployment and maintenance costs.
Quick Start
Follow the video tutorials below to quickly launch your first AI application.
Deploy a Chat-based Large Language Model
Deploy a Jupyter Notebook with PyTorch Environment
Detailed Documentation
For more usage guides, please refer to the following documents:
- How to use Chat-with-DeepSeek
- How to use Ollama (DeepSeek R1, Llama, Phi4, Gemma3, etc.)
- How to use ComfyUI
Template Details
Development Templates

Jupyter Minimal: A minimal Jupyter environment providing a basic runtime environment, allowing you to freely install required dependencies based on your needs.PyTorch: Built on jupyter-pytorch-notebook, pre-installed with PyTorch deep learning framework and JupyterLab, suitable for machine learning model development and training.TensorFlow: Built on jupyter-tensorflow-notebook, pre-installed with TensorFlow deep learning framework and JupyterLab, suitable for machine learning model development and training.R Notebook: Provides R language interpreter and basic data science environment, supporting R language development and data analysis.VSCode Server: A server-based Visual Studio Code environment accessible via web browsers, providing a complete code editing and development experience.
AI Models

DeepSeek R1: An open-source large language model (LLM), providing 1.5B, 8B, 14B, and OCR versions.Llama: An open-source large language model (LLM), providing Llama 3.2 in 3B and 11B versions.More Ollama models: Including Gemma3 4B, Qwen3 8B, and other open-source models.
Integrated Applications

PyTorch 2.7: Pre-installed with PyTorch 2.7 deep learning framework and JupyterLab, includes built-in Jupyter Notebook example documents for quick onboarding and development.Avatar generator: Integrated text-to-image model and modern web interface, supporting rapid generation of pixel-style or Studio Ghibli-style images and avatars.Digital Human: A digital human generation template based on ComfyUI workflows, integrated with IndexTTS and HeyGem, capable of generating video content from input text and images.ComfyUI with Notebook: Integrated ComfyUI and Jupyter Notebook with built-in common plugins, facilitating rapid workflow development and debugging.