sos1002/qwen-image-2-1-demo

Qwen-Image-2.1 Interactive Demo & ComfyUI Hub

An interactive showcase, prompt builder, and workflow toolkit for Comfy-Org/Qwen-Image-2.1 and Qwen/Qwen-Image-2.1.

🌟 Highlights of Qwen-Image-2.1

• —Native 2K Resolution: Up to 2688×1536 (16:9), 2048×2048 (1:1), 1536×2688 (9:16).
• —Superior Typography: State-of-the-art Chinese and English in-image text rendering.
• —Unified Text-to-Image & Image Editing: Multi-image inputs, style transfer, character consistency, and sequential storyboarding.
• —Transparent Output: Native alpha-channel (RGBA) background removal and graphic generation.
• —ComfyUI Ready: Official safetensors model weights packaged by Comfy-Org.

🛠️ ComfyUI Setup Guide

To run Qwen-Image-2.1 locally with ComfyUI:

1. 1.Diffusion Model: Download qwen_image_2.1_bf16.safetensors → ComfyUI/models/diffusion_models/
2. 2.Text Encoder: Download qwen3vl_8b_bf16.safetensors → ComfyUI/models/text_encoders/
3. 3.VAE: Download qwen_image_2.1_vae_bf16.safetensors → ComfyUI/models/vae/
4. 4.Load the official workflow templates included in the repository or generated directly from the web interface.

Created with Hugging Face Spaces & Antigravity.