<aside> 🛠

You are viewing a subpage of the CKNOOB User Guide, specifically the tutorial for the Chenkin Noob XL RF branch.

</aside>

<aside> 🛠

To view tutorials for the main CKNOOB model, please click CKNOOB User Guide.

</aside>

<aside> 😱

Machine Translation Warning: This document was translated using machine translation and may contain errors or omissions. The original text before translation is [here].

</aside>

<aside> 🛠

ChenkinNoob Authored by: 年糕特工队

</aside>

Introduction

The previous tutorial covered some features of CKNOOB, which will not be repeated here. This page focuses on explaining the CKNOOB RF model. Before formal use, the author will briefly introduce what RF prediction is and its benefits.

What is RF Forecasting?

RF Prediction (Rectified Flow) is an algorithm distinct from EPS and V-forecasting. Its core principle involves learning the direct path between data distribution and noise distribution, then continuously optimizing this path through iterative correction techniques. This enables efficient, rapid distribution transformation and sample generation.

Characteristics

Generally, CKNOOB RF exhibits the following characteristics compared to CKNOOB EPS:

• More stable composition
• Superior color rendering
• More refined brush strokes
• More tolerant of stroke counts while also supporting higher stroke counts.

Deployment Guide

Currently, it is recommended to use this model within ComfyUI.

Downloading the Model

You can download your desired model at [here].

Model Import

Move the downloaded model to the comfyui/models/checkpoints folder.