> ## Documentation Index
> Fetch the complete documentation index at: https://docs.monostate.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Interactive Wizard

> Train your first model with zero configuration

# Interactive Wizard

The AITraining Interactive Wizard guides you step-by-step through training your first AI model. No configuration files, no command memorization - just answer questions and start training.

## Why Use the Wizard?

<CardGroup cols={2}>
  <Card title="Zero Configuration" icon="wand-magic-sparkles">
    Answer simple questions instead of writing YAML files or memorizing CLI flags.
  </Card>

  <Card title="Smart Defaults" icon="brain">
    Intelligent defaults based on your hardware, model choice, and dataset size.
  </Card>

  <Card title="Validation Built-in" icon="shield-check">
    Catches errors before training starts - invalid datasets, missing columns, incompatible settings.
  </Card>

  <Card title="Learn As You Go" icon="graduation-cap">
    Every step has `:help` - learn what each option does while configuring.
  </Card>
</CardGroup>

## Launch the Wizard

```bash theme={null}
aitraining
```

That's it. Just run `aitraining` with no arguments and the wizard starts.

## What You'll See

```
 █████╗ ██╗████████╗██████╗  █████╗ ██╗███╗   ██╗██╗███╗   ██╗ ██████╗
██╔══██╗██║╚══██╔══╝██╔══██╗██╔══██╗██║████╗  ██║██║████╗  ██║██╔════╝
███████║██║   ██║   ██████╔╝███████║██║██╔██╗ ██║██║██╔██╗ ██║██║  ███╗
██╔══██║██║   ██║   ██╔══██╗██╔══██║██║██║╚██╗██║██║██║╚██╗██║██║   ██║
██║  ██║██║   ██║   ██║  ██║██║  ██║██║██║ ╚████║██║██║ ╚████║╚██████╔╝
╚═╝  ╚═╝╚═╝   ╚═╝   ╚═╝  ╚═╝╚═╝  ╚═╝╚═╝╚═╝  ╚═══╝╚═╝╚═╝  ╚═══╝ ╚═════╝
        From zero to hero Machine Learning Training Platform

🚀 Welcome to the AITraining Interactive Setup!

I'll help you configure your training job step by step.
Press Enter to accept defaults shown in [brackets].

Commands available at any prompt:
  :back  → return to the previous step
  :help  → show extra information for the current prompt
  :exit  → cancel setup
```

## Wizard Commands

These commands work at any prompt:

| Command           | What it Does                                |
| ----------------- | ------------------------------------------- |
| `:back`           | Go back to the previous step                |
| `:help`           | Show detailed help for the current question |
| `:exit`           | Cancel the wizard                           |
| `/search <query>` | Search for models or datasets               |
| `/sort`           | Change how models/datasets are sorted       |
| `/filter`         | Filter models by size (small/medium/large)  |
| `/refresh`        | Reload the model/dataset list               |

## Wizard Steps

The wizard guides you through these steps:

1. **Trainer Type** - What kind of AI? (LLM, classification, vision, etc.)
2. **Training Method** - For LLMs: SFT, DPO, ORPO, or PPO
3. **Project Name** - Where to save your trained model
4. **Model Selection** - Choose a base model (with hardware recommendations)
5. **Dataset Configuration** - Select and configure your training data
6. **Advanced Options** - Optional: hyperparameters, LoRA, logging

## Quick Start Paths

<Tabs>
  <Tab title="First Timer">
    **Just want to try it out?**

    1. Run `aitraining`
    2. Select "LLM" (option 1)
    3. Select "SFT" (option 1)
    4. Accept all defaults
    5. Watch it train!

    The wizard will pick a small model and sample dataset automatically.
  </Tab>

  <Tab title="Have Your Data">
    **Already have a dataset?**

    1. Run `aitraining`
    2. Choose your trainer type
    3. At the dataset step, enter your path: `./my_data` or `username/my-dataset`
    4. The wizard validates it and detects the format
    5. Configure columns if needed
  </Tab>

  <Tab title="Specific Model">
    **Want a specific model?**

    1. Run `aitraining`
    2. At the model step, use `/search llama` or `/search gemma`
    3. Use `/filter` to show only models that fit your hardware
    4. Select by number or type the full HuggingFace ID
  </Tab>
</Tabs>

## What's Next?

<CardGroup cols={2}>
  <Card title="SFT Training Walkthrough" href="/wizard/sft-walkthrough" icon="play">
    Complete step-by-step guide for training your first LLM
  </Card>

  <Card title="Understanding Models" href="/wizard/choosing-models" icon="robot">
    How to pick the right model for your hardware and use case
  </Card>

  <Card title="Dataset Guide" href="/wizard/dataset-guide" icon="database">
    Understanding datasets, formats, and what works best
  </Card>

  <Card title="Commands Reference" href="/wizard/commands" icon="terminal">
    All wizard commands and shortcuts
  </Card>
</CardGroup>