FireRed Image Edit in ComfyUI | Qwen Image Edit Workflow, Multi-Reference Edits & Restoration Tests

Описание: In this video, we explore FireRed Image Edit, a fine-tuned model
based on Qwen Image Edit, and test its capabilities inside a
complete ComfyUI workflow.

This walkthrough covers:

Model setup (BF16, FP8, and GGUF options)
Lightning 4-step LoRA integration for faster sampling
Multi-reference image editing
Old photo restoration and colorization
Pattern transfer using reference images
Object insertion (hat transfer test)
Localized edits (color change without affecting the entire scene)
Custom resolution outputs (1280x720 test)

We test FireRed Image Edit across multiple real-world scenarios
including vintage photo restoration, identity preservation, texture
transfer using a tiger reference image, and localized jacket color
edits.

Since FireRed is based on Qwen Image Edit, it supports compatible LoRAs
including lighting LoRAs, making it a flexible option for creators using
ComfyUI.

🔹 Is FireRed better than Qwen Image Edit?
🔹 Does it preserve identity well?
🔹 How does it handle multi-image compositing?

Watch the full video to see the results.

------------------------------------------------------------------------

Models Used

FireRed Image Edit
Qwen Image Edit 25-11 Lightning 4-Step LoRA
Qwen 2.5 7B VL (Text Encoder)
Qwen Image VAE

------------------------------------------------------------------------

🔹 Download Links:

diffusion_models

BF16: https://huggingface.co/cocorang/FireR...
FP8: https://huggingface.co/cocorang/FireR...
FP8 (mixed): https://huggingface.co/cocorang/FireR...

GGUF

https://huggingface.co/vantagewithai/...

LoRAs

https://huggingface.co/lightx2v/Qwen-...

text_encoders

https://huggingface.co/Comfy-Org/Huny...

vae

https://huggingface.co/Comfy-Org/Qwen...

workflow

https://huggingface.co/vantagewithai/...

Model Storage Location

📂 ComfyUI/
├── 📂 models/
│ ├── 📂 text_encoders/
│ │ └── qwen_2.5_vl_7b_fp8_scaled.safetensors
│ ├── 📂 loras/
│ │ └── Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors
│ ├── 📂 diffusion_models/
│ │ └── FireRed-Image-Edit-1.0_bf16.safetensors
│ ├── 📂 unet/
│ │ └── FireRed-Image-Edit-1.0-XX-X.gguf
│ └── 📂 vae/
│ └── qwen_image_vae.safetensors



If you found this guide helpful, consider liking the video and
subscribing for Part 2, where we directly compare FireRed Image Edit
against Qwen Image Edit and push the limits further with advanced
lighting and identity tests.

#ComfyUI #QwenImageEdit #FireRedImageEdit #AIImageEditing