Google Releases a Free Photo AI Enhancement and Variants for Its Gemma Open-Source Model

IBL News | New York

Google announced this week that it will bring for free to users its Photos application AI-powered editing features, which were previously limited to Pixel phones and tablets and paid subscribers. This enhancement, which will start rolling out on May 15, includes Magic Editor, Magic Eraser, Photo Unblur, and Portrait Light.

Magic Editor, the most notable feature, uses generative AI to do more complicated photo edits, like filling in gaps in a photo and repositioning the subject to the foreground or background.

On the other hand, Google announced that the Gemma family of lightweight, open-source AI models will be extended with two new variants: CodeGemma for code generation, and RecurrentGemma, designed to improve inference at higher batch sizes, which is useful for researchers.

CodeGemma models are available as a 7B pre-trained variant that specializes in code completion and code generation tasks, a 7B instruction-tuned variant for code chat and instruction-following, and a 2B pre-trained variant for fast code completion.

Finally, Google unveiled that will make a new chip for AI work, named Axion, trying to combat the dominance of NVIDIA and other large companies.