[ad_1]
Google recently announced the Gemini AI along with Imagen 2 and Duet AU for developers. These can be used to develop tools and integrate them into their products and services. These are all available with Gemini, Imagen 2 and Duet AI on Vertex AI (Google Cloud API solution via AI Studio).
The company is rapidly expanding the generative model of their AI and the latest APIs. There’s no denying that Google has improved the quality of its software while ensuring it remains reliable and cost-effective.
Developers can access Gemini, Imagen 2, and Duet AI on Vertex AI, where they get the IDE to test and develop applications with the Google GenAI AI models. Companies can use Vertex AI with the available AI models to build foundational models using their own data sets.
Developers can use the API on Vertex AI to develop applications.
Vertex AI offers more than 130 Google AI models to help companies improve the quality and accuracy of their search applications.
The company’s most advanced and capable multimodal model can process information including text, code, images and videos.
- Based on the requirements, developers can adjust the parameters. It also has the ability to improve through RLHF based on user feedback.
- It is also useful for adapting to specific contexts or use cases. Gemini Pro can be linked to an external API to retrieve data sources.
- It also ensures scalability with tools that make deployment and maintenance easier.
- It has the ability to create and power search, summarize, and answer generation functions.
- Voice and chat agents are also available in preview.
Developers can also access Gemini Pro through Google AI Studio, a web-based cue testing tool.
Image 2
This is Google’s AI text-to-image generation model, Imagen 2, which is said to offer major improvements in image quality and new features.
- Ability to generate photorealistic, attractive, high-resolution images using natural language prompts.
- It now supports multiple languages to create images with accurate text overlays.
- Now generating logos has become extremely easy for products or companies, and they are also overlaid on an image.
- Can analyze the image to provide informative text responses to questions about image details.
Currently, it is available to limited users like Snap by offering Imagen with AI Camera Mode to Snapchat+ subscribers. This allows users to choose from pre-selected options to generate scenes. It is also joined by Shutterstock, which launched AI image generators using text prompts. However, Shutterstock has trained its AI image generators on more than 16,000 images.
Duet AI
Duet AI for developers and security operations has been upgraded to Gemini Pro. Now it’s generally available in Google Workspace.
- It supports more than 20 programming languages, including C, C++, Go, Java, JavaScript and Python.
- IDEs add several integrations such as Cloud Shell Editor, Cloud Workstations, IntelliJ, PyCharm, and Visual Studio Code.
- Context-aware latency is reduced. Now Duet AI generates content in natural language chat faster.
As it now offers AI-powered code and chat assistance, it also plans to customize Duet AI for developers, unlocking more possibilities.
Google says they won’t use developers’ data to train their AI model. It remains under the control of the users. Google’s main release is scheduled for early next year, Ultra, which is currently available to select users before it becomes available to everyone.