In response to the burgeoning generative AI landscape, Google has announced a series of significant enhancements to its Vertex AI platform. The are designed to equip developers with advanced tools to navigate the rapidly evolving world of generative AI applications.
Vertex AI is a machine learning platform that lets users train, deploy, and customize models and applications for various data types and modalities. It combines data engineering, data science, and ML engineering workflows, enabling collaboration and scalability using Google Cloud.
Vertex AI provides several options for model training and deployment, such as AutoML, custom training, model garden, and generative AI. It also offers end-to-end MLOps tools to automate and scale projects throughout the Machine Learning lifecycle, such as workbench, vizier, experiments, model registry, feature store, explainable AI, and model monitoring.
Key Announcements from Google Cloud Next
During its Google Cloud Next conference, Google unveiled many advancements aimed at bolstering the capabilities of Vertex AI. The official announcement on Google's Cloud Blog provided a comprehensive breakdown of these enhancements.
Google's Model Garden is set to witness the inclusion of several new models. Other noteworthy additions include Llama 2 and Code Llama from Meta, the Technology Innovation Institute's Falcon LLM, and a forthcoming model named Anthropic's Claude 2. These models represent a curated collection spanning first-party, open-source, and third-party offerings.
Upgrades to First-Party Foundation Models
Google's proprietary foundation models have also undergone pivotal upgrades. The PaLM model now offers enhanced output quality and an extended 32,000-token context window, simplifying the analysis of voluminous documents. Meanwhile, Codey, tailored for code generation and chat functionalities, boasts improved performance metrics. The image generation model, Imagen, has been refined to deliver superior image quality.
Digital Watermarking with DeepMind Collaboration
In a strategic collaboration with Google DeepMind, Google introduced a novel digital watermarking feature for Imagen. Powered by DeepMind's SynthID, this feature facilitates the verification of AI-generated images, ensuring authenticity and credibility.
Real-time Data Extensions
The launch of Vertex AI Extensions marks a pivotal step in connecting models to real-time data via APIs. These extensions are versatile, catering to the development of digital assistants, search engines, and intricate automated workflows.
Customer Testimonials Highlight Value
The official announcement was punctuated with testimonials from satisfied customers, emphasizing the transformative impact of these Vertex AI enhancements. These endorsements underscore the platform's capability to revolutionize workflows, enhance software development, and drive content personalization at an unprecedented scale.