A beginner's guide to the T2i-Adapter-Sdxl-Lineart model by Adirik on Replicate
Published: (January 4, 2026 at 09:50 PM EST)
1 min read
Source: Dev.to
Source: Dev.to
Overview
The T2I-Adapter-SDXL‑Lineart model is a text‑to‑image generation model developed by Tencent ARC. It extends Stable Diffusion XL (SDXL) by adding line‑art conditioning, allowing the model to modify or generate images based on both a textual prompt and a line‑art input.
Model family
T2I‑Adapter‑SDXL‑Lineart belongs to a family of adapters created by Tencent ARC:
- t2i‑adapter‑sdxl‑sketch – uses sketch images as conditioning.
- masactrl‑sdxl – provides editable image generation capabilities.
All these models share the same underlying SDXL checkpoint but differ in the type of visual conditioning they accept.
Usage parameters
| Parameter | Description |
|---|---|
| Image | The input line‑art image that serves as conditioning for generation. |
| Prompt | Text describing the desired output image. |
| Scheduler | Diffusion scheduling algorithm (default: K_EULER_ANCESTRAL). |
| Num Samples | Number of images to generate (maximum 4). |
| Random Seed | Optional seed for reproducible results. |
| Guidance Scale | Controls how strongly the output follows the text prompt. |
| Negative Prompt | Text specifying elements that should be avoided in the output. |
| Num Inference Steps | Number of diffusion steps (up to 100). |
| Adapter Conditioning Scale | Determines the influence of the line‑art conditioning. |
| Adapter Conditioning Factor | Controls the overall size of the generated image. |
Output
The model returns an array of generated images as image URIs.