Google Cloud brings the text-to-live image feature to Imagen & other updates for Gemini, Gemma, and MLOps

All the popular tech companies are working on their AI technology to improve user experiences. On April 10, 2024, Google Cloud announced model updates for Gemini, Gemma, Imagen, and platform capabilities that continue to enhance Vertex AI.

Imagen, a task-specific generative AI model, has become an important tool for companies that leverage AI-driven creativity at a large scale. It has received significant updates. Let’s discuss them.

Text-to-Live image capabilities

One of the standout features discussed was the preview of Imagen 2.0’s text-to-live image capabilities. This feature empowers creative and marketing teams to transform text prompts into animated images, like GIFs.

Google Cloud mentioned that live images will be delivered at 24 frames per second with a resolution of 360*640 pixels, and the time taken to generate the image will be 4 seconds. The search giant assured us that they plan to improve the quality and reduce the time taken by providing further updates.

Focused design of enterprise apps

Imagen 2.0’s design emphasizes its suitability for enterprise applications, with its focus on delivering extraordinary results in themes like food imagery, nature, and animals.

The model is capable of generating a diverse range of camera angles and motions successfully while maintaining consistency throughout.

To maintain the integrity and authenticity of the generated content, Imagen 2.0 for live image generation also comes equipped with powerful safety filters and digital watermarks.

Advanced photo editing features

Along with text-to-live image capabilities, Imagen 2.0 also comes with photo editing features, including inpainting and outpainting. These will allow you to remove unnecessary elements from your image and add new elements, as well as expand image borders to create a wider field of view.

The digital watermarking feature, powered by Google DeepMind’s SynthID, lets you generate invisible watermarks, thereby providing a layer of security and verification for images and live images generated by the Imagen family of models.

Big companies like Shutterstock and Rakuten are already using Imagen 2.0 to create high-quality, accurate images at the enterprise level.

The text-to-live image capabilities will expand Imagen’s potential applications, allowing companies across different industries to unlock new avenues of creativity.

Other announcements

Furthermore, Google Cloud’s Vertex AI platform introduced a suite of updates to revolutionize AI development and deployment. Gemini 1.5 Pro is now available in public preview, allowing developers to access the world’s largest context window, facilitating multimodal reasoning over vast amounts of data.

This capability can be used in diverse fields, from customer service automation to financial document analysis

With the expansion of MLOps capabilities for gen AI, including new prompt management and evaluation services for large models, organizations can manage and deploy models in production, thereby streamlining the transition from experimentation to scale.

Vertex AI also offers integration with Google Search, which will enhance response accuracy, providing a better user experience while surfing the web.

In the blog post, Amin Vahdat, VP/GM ML, Systems, and Cloud AI, Google also mentioned the expansion of data residency guarantees and said:

Today we’re also expanding data residency guarantees — which cover data stored at-rest for Gemini, Imagen, and Embeddings APIs on Vertex AI — to 11 new countries: Australia, Brazil, Finland, Hong Kong, India, Israel, Italy, Poland, Spain, Switzerland, and Taiwan. Customers can now also limit machine learning processing to the United States or European Union when using Gemini 1.0 Pro and Imagen.

In conclusion, these announcements about advancements in AI tech via Vertex AI and Imagen AI show a significant leap forward with the aim of empowering enterprises to leverage the potential of generative AI.

What are your thoughts on the matter? Share your opinions with our readers in the comments section below.