Alibaba's Qwen Opensources Image Model Qwen-Image-2.1 with 7B Parameters for Transparent Image Generation and Editing

Alibaba's Qwen team has announced the open-source release of its image model Qwen-Image-2.1, which combines text-to-image generation and image editing in a single model. The visual generation part has only 7B parameters and natively supports transparent image generation and editing.

According to the team, the new model has four key features: a lightweight structure that balances generation quality and computing cost; support for generating regular or transparent images from prompts, with transparent layer editing and cutout; support for up to 10 reference images to enhance local editing, portrait and product fidelity; and improved text layout, lighting and detail rendering.

The model is now available on GitHub, HuggingFace and ModelScope for developers to download.