Quick Start Guide¶

Welcome to DeepExtension — a unified platform for domain-specific LLM training, evaluation, and deployment.

This quick guide will help you get started — from installation to your first training run — using our no-code, web-based UI.

Prerequisite: Installation First¶

Before you can start using DeepExtension, you need to set up the platform. Installation typically requires some development experience.

Refer to the full Installation Guide for step-by-step instructions.
You can install DeepExtension on:
- Linux or Windows via WSL (with CUDA for GPU training)
- macOS (with MLX on Apple M-series)
- Any Linux/macOS environment (in no-training mode) for UI exploration and inference only

Once the installation is complete and the Web UI is accessible (default: http://localhost:88), you're ready to move forward.

Your First Training Run¶

After installation, we recommend starting with a demo training method. DeepExtension includes preinstalled training methods depending on your platform.

See Quick Start: Run Your First Training for hands-on instructions.

You'll be guided through:

Uploading a sample dataset
Choosing a base model
Launching a training job

All demo training methods come with sample datasets and configurations so you can run a successful training in minutes.

After Training¶

Once your training job completes successfully, you can:

Explore model results in the Customized Models tab
Save or deploy your model for inference
Run prompt-based evaluations using DeepExtend
Compare models using the Model Assessment module

Where to Go Next¶

Tutorial Overview — A complete map of prerequisites and training flows
User Guide — Learn each module in detail
Developer Guide — If you're customizing training or contributing your own model logic

DeepExtension — Extend AI capabilities for your business.