Skip to content
WorkflowListedExternal
Open source

Qwen-image-workflow_gguf_Lowvram — 4/8step

Lightx2v is awesome, seriously awesome! 4-step LoRA is amazing. Now, using this workflow, we can get results in 8 steps that are as good as the 20-step workflow.

Sourcecivitai.com/models/1841922
TaskText generationFamilyQwen-ImageEstimated VRAMNot yet characterizedBgrade

Runtime profile

Access
External workflow
Status
Listed
Source
Civitai
Updated
Oct 2, 2026

Source description

4step:

Lightx2v is awesome, seriously awesome! 4-step LoRA is amazing. Now, using this workflow, we can get results in 8 steps that are as good as the 20-step workflow.

Adding shift nodes, It makes a huge difference to the image quality when using the Lighting LoRA.

https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-4steps-V1.0.safetensors

8step:

We can shave off 60% of the generation time now.

https://huggingface.co/lightx2v/Qwen-Image-Lightning/blob/main/Qwen-Image-Lightning-8steps-V1.0.safetensors

V1:

This is the basic workflow for Qwen-Image.

Qwen-Image is good at complex text generation and prompt-following . You can even use it to create intricate text covers, memes, etc.

Official examples: https://qwenlm.github.io/blog/qwen-image/

Glad to see city96's fresh Qwen-Image-gguf checkpoint has arrived!

Checkpoint gguf https://huggingface.co/city96/Qwen-Image-gguf/tree/main "

What you need

Model files
29.9 GiB known subtotal

Remaining hardware or download-size estimates are not available yet.

Requirements

Models

8

  • ae.safetensors

    File unverified

    VAE · Hugging Face · Comfy-Org/Lumina_Image_2.0_Repackaged · Lumina Image 2.0 Repackaged

  • CLIPLoader

    Possible

    Text encoder · Unknown

  • qwen_2.5_vl_7b_fp8_scaled.safetensors

    Found

    Text encoder · 8.74 GB · SAFETENSORS · fp8 · pruned · Unknown

  • qwen_2.5_vl_fp16.safetensors

    File unverified

    Text encoder · Hugging Face · Comfy-Org/Omnigen2_ComfyUI_repackaged

  • qwen_image_vae.safetensors

    Possible

    VAE · Unknown

  • qwen-image-Q4_K_S.gguf

    Found

    Checkpoint · 11.3 GB · GGUF · Unknown

  • VAEDecode

    Possible

    VAE · Unknown

  • VAELoader

    Possible

    VAE · Unknown

Node packs

9

  • CFGZeroStar

    Not resolved

    Node pack · Unknown

  • CLIPTextEncode

    Possible

    Node pack · Unknown

  • EmptySD3LatentImage

    Not resolved

    Node pack · Unknown

  • KSampler

    Possible

    Node pack · Unknown

  • ModelSamplingAuraFlow

    Found

    Node pack · Unknown

  • Power Lora Loader (rgthree)

    Not resolved

    Node pack · Unknown

  • SaveImage

    Possible

    Node pack · Unknown

  • UnetLoaderGGUF

    Possible

    Node pack · Unknown

  • UNetTemporalAttentionMultiply

    Not resolved

    Node pack · Unknown

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • Use Qwen-image-workflow_gguf_Lowvram to create images from text prompts, including intricate text covers and memes; its description highlights complex text generation and prompt-following.

Inputs and outputs

  • The described input is a text prompt; examples of generated images include text covers and memes.

Setup

  • Before setup, confirm which qwen_image_vae.safetensors file to use; only the filename is listed.

    Suggestion · not verified

Dependencies

  • Named model files include ae.safetensors, qwen_2.5_vl_7b_fp8_scaled.safetensors, qwen_2.5_vl_fp16.safetensors, qwen_image_vae.safetensors, and qwen-image-Q4_K_S.gguf.

  • Listed node names include CFGZeroStar, CLIPLoader, CLIPTextEncode, EmptySD3LatentImage, KSampler, and ModelSamplingAuraFlow.

  • Other listed node names are Power Lora Loader (rgthree), SaveImage, UnetLoaderGGUF, UNetTemporalAttentionMultiply, VAEDecode, and VAELoader.

Configuration

  • The description links Qwen-Image-Lightning-4steps-V1.0.safetensors and Qwen-Image-Lightning-8steps-V1.0.safetensors.

  • When using the Lighting LoRA, the author suggests adding shift nodes.

    Suggestion · not verified

Cautions

  • The Civitai listing does not state a usage license.

  • Check the listing's terms before using the workflow or sharing generated images.

    Suggestion

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimary
civitai.com/models/1841922

Verified Jul 19

Source excerpts

1 excerpt
NoteJul 19civitai.com/models/1841922

Source context: 1395 downloads · Type Workflows · Base model Qwen