# SDXL

SDXL is a text-to-image diffusion model capable of generating photo-realistic images given any text input. You can learn more about the model [here](https://replicate.com/stability-ai/sdxl).

You can generate images using SDXL by using the "Create Prediction - SDXL" workflow action. 

It requires the following inputs:

<table><thead><tr><th width="154">Input</th><th width="145.99999999999997">Type of Field</th><th>Description</th></tr></thead><tbody><tr><td>Prompt</td><td>Text</td><td>The text prompt you're using to create the image(s)</td></tr><tr><td>Image</td><td>image</td><td><p></p><p>Input image for img2img or inpaint mode</p><p></p></td></tr><tr><td>Mask</td><td>mask</td><td>Input mask for inpaint mode. Black areas will be preserved, white areas will be inpainted.</td></tr><tr><td>Height</td><td>Number (integer)</td><td>The height of the image(s) you want to create</td></tr><tr><td>Width</td><td>Number (integer)</td><td>The width of the image(s) you want to create</td></tr><tr><td>Number of Outputs</td><td>Number (integer)</td><td>The number of images you want to create with each prediction. Must be between 1 and 4.</td></tr><tr><td>Negative Prompt</td><td>Text</td><td>Things you do NOT want to see in the output image(s)</td></tr><tr><td># Inference Steps</td><td>Number (integer)</td><td>The number of denoising steps used in the prediction (minimum: 1; maximum: 500). In general, the more steps you use the more detailed the output. However, the more steps you use the longer it will take to generate the prediction.<br></td></tr><tr><td>Guidance Scale</td><td>Number (decimal)</td><td>Scale for classifier-free guidance (minimum: 1; maximum: 20)</td></tr><tr><td>Seed</td><td>Number (integer)</td><td>Random seed. Leave blank to randomize the seed</td></tr><tr><td>Scheduler</td><td>Text (list of pre-defined options)</td><td>The type of scheduler you want to use.</td></tr><tr><td>Prompt Strength</td><td>Number (decimal) between 0 and 1.0</td><td>Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image</td></tr></tbody></table>