# Text to Image With ControlNet

Generate images based on depth or normal information obtained from the Unity camera

| Input                            | Depth                            | Normal                           |
| -------------------------------- | -------------------------------- | -------------------------------- |
| ![](/files/7MnTCGk4WvP7W8BaLAws) | ![](/files/uq9PYWav9nImi5VAhTMK) | ![](/files/II5an6MptgqIhihw8tSL) |

This makes it possible to use image generation in areas related to the game system where geometry is important. 

Note: ControlNet requires a large amount of VRAM. If you are unable to generate images due to insufficient VRAM, you may need to reduce the resolution.<br>

<figure><img src="/files/tPndbbAZPgbZ4mGlRhTB" alt=""><figcaption></figcaption></figure>

For generate image with Depth to Image, attach the Text2ImgWithControlNet.cs component to the object

Uncheck the CullingMask UI checkbox.&#x20;

If you do not do this, the UI element will also get the depth information of the nearest neighbor, so it will be affected by the following RawImage, etc.

<figure><img src="/files/L9VrMbXetmJxDI5fFoLl" alt=""><figcaption></figcaption></figure>

It is assumed that the local server is up and running.&#x20;

Please refer to the following for StartServer before use.

{% content-ref url="/pages/lMqSgGDhqUVLFoLqfYMZ" %}
[About local server](/english/production-tips/about-local-server.md)
{% endcontent-ref %}

Set the base 3D models in the range of the camera

<figure><img src="/files/Ui80UINZYq6xp7S8BNHX" alt=""><figcaption><p>camera range</p></figcaption></figure>

<figure><img src="/files/OosyUG8Klx9E1Cnuetu8" alt=""><figcaption><p>place 3D model in the range of camera</p></figcaption></figure>

<figure><img src="/files/wK8WzGP7mez7MaKCXayV" alt=""><figcaption></figcaption></figure>

A parameter summary is shown below. Parameters that are adjusted frequently are bolded.

<table><thead><tr><th width="177">Parameter</th><th>Summary</th></tr></thead><tbody><tr><td><strong>Prompt</strong></td><td>Enter the words you want to generate in natural language</td></tr><tr><td><strong>NegativePrompt</strong></td><td>Input elements in natural language that you do not want to include in the generated image</td></tr><tr><td><strong>Size</strong></td><td>Width and hight of the generated image in pixels</td></tr><tr><td>Steps</td><td>Increase this parameter if the generated image looks blurry</td></tr><tr><td><strong>Cfg Scale</strong></td><td>Strength to follow prompts<br>Input larger value to strongly follow the prompt.</td></tr><tr><td>Seed</td><td><p>Fixing the seed value to generate a similar image</p><p>In case of -1, the image will be random</p></td></tr><tr><td><strong>Batch Count</strong></td><td><p>The number of images to be generated.</p><p>Select 1 to 100 images.</p><p>Generated images will be stored in StreamingAssets/StableDiffusion. Please check images by date sort in the explorer.</p></td></tr><tr><td><strong>Camera</strong></td><td>Set the camera with setting up for getting depth</td></tr><tr><td><strong>ControlType</strong></td><td>You can choose Depth or Normal</td></tr><tr><td><strong>Weight</strong></td><td>You can adjust the intensity of ControlNet's influence.</td></tr><tr><td>Sampler</td><td>Changing the nuances of the generated image</td></tr><tr><td><strong>Model</strong></td><td>Base trained model<br>See the "<a href="/pages/kXnr6Qlq0viYovRdupEd">Switch Models/LoRA</a>" to switch new model</td></tr><tr><td><strong>Lora</strong></td><td><p>Additional trained Lora to generate consitent specific pattern image</p><p>See the "<a href="/pages/kXnr6Qlq0viYovRdupEd">Switch Models/LoRA</a>" to switch new LoRA</p><p>If you select Lora, &#x3C;lora:filename:1>will be added to the Prompt and LoRA will be available</p></td></tr></tbody></table>


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.witchpot.com/english/components/text-to-image-with-controlnet.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.