1. Vertex AI offers access to a wide variety of foundation models, robust infrastructure options, and tools for model development and MLOps.
2. Updates introduced at Google Cloud Next include Gemini 1.5 Pro and Imagen 2.0 enhancements for generative AI app development.
3. Additional features like grounding responses with Google Search, prompt management services, and evaluation services for large models are expanding MLOps capabilities for gen AI on Vertex AI.
Vertex AI is a comprehensive platform for building, deploying, and maintaining generative AI apps and agents, offering a wide variety of foundation models, infrastructure options, and tools for model development and MLOps. New updates and capabilities have been introduced at Google Cloud Next, including Gemini 1.5 Pro, Imagen 2.0 for image generation, and CodeGemma.
Enhancements have been made to response accuracy in gen AI services within Vertex AI, with the ability to ground responses with Google Search now in public preview. This provides users with access to fresh, high-quality information that improves the overall accuracy of model responses.
MLOps capabilities for gen AI have been expanded in Vertex AI to help customers manage and deploy models effectively in production, with new prompt management and evaluation services for large models. These features enable organizations to optimize the performance of gen AI models at scale and iterate quickly from experimentation to production.
Gemini 1.5 Pro, now in public preview on Vertex AI, offers the world’s largest context window for developers, enabling multimodal reasoning over vast amounts of data specific to a request. Use cases for this model include customer service agents, academic tutors, financial document analysis, and more. Imagen 2.0 is also featured, allowing for the creation of short live images from text prompts with capabilities for image editing, inpainting/outpainting, and digital watermarking.