Qwen-Image-Edit-2511: Comprehensive Image Editing Workflow
Details
Download Files
Model description
You can click on the link below to try it out directly. If the effect is good, you can deploy it locally
https://www.runninghub.ai/post/2003529388635398145/?inviteCode=sdhs0trb
Fan benefits,register to get 1000 points,daily login 100 points,play 4090!Experience the super power of 48G.
Qwen-Image-Edit-2511 FP8 Usage Guide
Qwen-Image-Edit-2511 is the latest image editing model released by the Alibaba Qwen team. Compared to previous versions, it offers significant improvements in character consistency, multi-person editing, and instruction following.
1. Quick Start Guide
To use this model in ComfyUI or similar workflows, follow these steps:
Model Installation:
Place the downloaded FP8 Main Model (e.g.,
qwen_image_edit_2511_fp8_e4m3fn.safetensors) into yourComfyUI/models/diffusion_models/directory.Ensure you have the compatible CLIP/Text Encoder (recommended:
qwen_2.5_vl_7b_fp8_scaled) and VAE models installed.
Key Features:
Instruction-Based Editing: Modify images using natural language (e.g., "Change the character's outfit to red" or "Add a sunset to the background").
Character Consistency: Perfectly maintains original facial features while modifying environments or actions.
High-Speed Sampling: When paired with Lightning LoRA, it generates high-quality results in just 4–8 steps.
Operational Tips:
Input Image: Upload the source image you wish to edit.
Prompt: Enter clear and concise editing instructions.
CFG Scale: For FP8 models, keep CFG between 1.0 - 1.5 (when using Lightning LoRA) or follow the default workflow settings.
2. Hugging Face Download Links
Recommended FP8 Quantized Version (by xms991): https://huggingface.co/xms991/Qwen-Image-Edit-2511-fp8-e4m3fn
Official Model Weights (BF16): https://huggingface.co/Qwen/Qwen-Image-Edit-2511
Lightning Optimized Version (by lightx2v): https://huggingface.co/lightx2v/Qwen-Image-Edit-2511-Lightning/tree/main