Qwen Image Edit: Direct characters with prompts

Qwen Image Edit is an image-editing model. It enables you to make precise edits to images, from subtle retouches to complex scene transformations, all through simple text prompts.

Qwen Image Edit can preserve lighting and style while seamlessly integrating your edits. It supports both edits using text prompts alone and with masks (inpainting).

For example, you can use Qwen Image Edit to create different viewing angles of the same character.

In this tutorial, you’ll learn how to:

• Set up Qwen Image Edit on ComfyUI or Google Colab
• Use prompt-based editing
• Rotate a character and cast her to a different scene
• Export and upscale your final image for production use

Let’s dive in and explore how to edit images intelligently with Qwen!

Software

ComfyUI Colab Notebook

If you don’t have a powerful GPU card, you can still run the Qwen Image model on Google Colab, Google’s cloud computing platform, with my ComfyUI notebook.

You don’t need to download the model as instructed below. Select the Qwen_Image_Edit model before running the notebook.

Use cases

To showcase the power of Qwen Image Edit, here are some edits generated using the prompt-only workflow.

Rotate a character

Front view

back view

Changing background

You can change the background easily with Qwen Image Edit.

She’s in New York City

Adding objects

You can ask her to hold any object!

she’s holding a light saber on a spaceship

she’s holding a pot of small cactus on a private ship

she’s wearing business formal dress in gray and white. keep the same background

Qwen Image Edit (Prompt)

This workflow edits an image using instructions in the prompt alone.

This workflow uses Qwen Image Lightning LoRA to speed up the workflow to 4 steps.

Step 0: Update ComfyUI

Before loading the workflow, make sure your ComfyUI is up to date. The easiest way to do this is to use ComfyUI Manager.

Click the Manager button on the top toolbar.

Select Update ComfyUI.

Restart ComfyUI.

Step 1: Install models

After loading the workflow JSON file, ComfyUI should prompt you to download the missing model files. You can use the dialog box to download the models, or use the links below.

Here are the models you need to download:

• Download qwen_image_edit_fp8_e4m3fn.safetensors and put it in ComfyUI > models > diffusion_models.
• Download qwen_2.5_vl_7b_fp8_scaled.safetensors and put it in ComfyUI > models > text_encoders.
• Download qwen_image_vae.safetensors and put it in ComfyUI > models > vae.
• Download Qwen-Image-Lightning-4steps-V1.0.safetensors and put it in ComfyUI > models > loras.

Step 2: Load the workflow

Download the workflow below. Drop it into ComfyUI to load.

Step 3: Upload an image to edit

Upload the image you wish to edit.

You can use the test image below.

Step 4: Revise the prompt

Write the instructions for editing in the prompt.

Step 5: Generate an image

Click the Run button to run the workflow.

Run it again to generate a variation.

Qwen Image Edit Upscale workflow

The input image is rescaled to a fixed size so that the Qwen Image Edit model can process it. However, you can enlarge the final image using an AI upscaler.

Workflow

Use the following nodes to upscale the final image.

• Uscale image (using model)
• Load Upscale Model

Here’s the Qwen Image Edit with upscaling nodes:

Install Upscale Model

You will need an upscale model to run this workflow.

Download the 4x-Ultrasharp model. Put it in ComfyUI > models > upscale_models folder.

Running the workflow

Follow the same instructions as the last section to run this workflow.

The upscaled image is on the left in the Upscale group.

Next Steps

In the next post, I will show you how to use Qwen Image Edit with masks, i.e., inpainting. Stay tuned!