South Korea's Kakao Brain, a subsidiary of Kakao Corp., has released its new AI model, Karlo 2.0.
The AI model can generate highly detailed images in as little as three seconds, a feature attributed to its training on a dataset of approximately 300 million text and image pairs.
Why it matters: AI image generation technology is critical in many fields, including art, gaming, and virtual reality.
Kakao Brain's new release could speed up processes and improve results in these fields, with implications for the broader AI ecosystem.
The Key Points
- Karlo 2.0 can interpret complex instructions such as "A cat has light blue eyes" and generate corresponding images.
- The model supports resolutions up to 2048x2048 pixels, providing flexibility to work with different image sizes.
- Kakao Brain has accelerated the image creation process with Karlo 2.0, reducing the time to three seconds.
- The company also released the Karlo 2.0 API to Kakao Developers, offering up to 600,000 free image generation to developers and startups monthly.
The Big Picture: The launch of Karlo 2.0 is part of Kakao Brain's continuous efforts to advance its AI technology and support the development of the domestic AI ecosystem.
As the new model has improved image generation capabilities, it could significantly impact AI-generated visuals and provide more opportunities for developers and businesses.
According to Kim Ildoo, CEO of Kakao Brain, the company aims to create photorealistic images and develop a model that meets a broader range of user needs.