Today's selected cutting-edge AI information, welcome to read 👇

🖼️ Meta AI releases SAM 2.1 image and video segmentation model, with comprehensive performance improvements, and open-sources training and demo code.

🎨 Coframe collaborates with OpenAI to develop UI code generation model, will fine-tune GPT-4o vision model to improve UI generation quality.

📝 Text Behind Image: Open-source tool that can add text behind specified objects in images, supporting various scenarios and local saving.

Cutting-edge News

1. Meta AI releases image and video segmentation model SAM 2.1.

From the various scores provided, there are improvements compared to SAM 2. They also released training code and a set of frontend + backend web demo code.

Detailed introduction: https://github.com/facebookresearch/sam2?tab=readme-ov-file#latest-updates

2. Coframe announces collaboration with OpenAI to develop UI code generation model.

Currently, large language models struggle to generate high-quality UI interfaces. This time, Coframe is collaborating with OpenAI to fine-tune the GPT-4o vision model, using a large amount of UI data for training.

Detailed introduction: https://www.coframe.com/post/coframe-works-openai-to-develop-first-of-its-kind-ui-code-generation

The fine-tuned model will be much better in generating aesthetically pleasing and visually appealing UIs, allowing us to quickly and easily build striking sections for websites.

Open Source Projects

1. A simple but quite interesting open-source free tool: Text Behind Image.

It can easily add text behind specified objects in images, such as animals, humans, objects, etc., and supports saving to local storage.

GitHub: https://github.com/RexanWONG/text-behind-image