The Oracle Cloud Infrastructure now features the Generative AI service. This makes it easier for apps on OCI to use large language models and generative AI.
The new service offers LLMs from Cohere and Meta for this purpose. Meta’s Llama-2 is best suited for text generation, while Cohere offers the three models “Command,” “Summarize,” and “Embed”. Command is suitable for text generation, Summarize is more trained to generate qualitative summaries by extracting critical information from documents, and Embed converts text to vector embeddings.
Customize models
Users of the new Oracle service can integrate these models into their apps and refine them with their data. Oracle supports this process with retrieval augmented generation (RAG) techniques, allowing the models to understand internal tasks and business processes better.
Oracle is offering the OCI Generative AI Agents service in beta to support this. With an RAG agent, LLMs can be used with enterprise search on OCI Open Search to provide context to search results with data from an enterprise. The agent supports multiple data sources and must always be up-to-date.
Initially, Oracle only supports OCI OpenSearch, but other tools will also receive support later. Oracle will also introduce preconfigured agent actions for its software applications, such as the Fusion Cloud Applications Suite, NetSuite and Cerner.
This launch is part of Oracle’s broader AI offerings. “By integrating generative AI across its portfolio of cloud applications—including ERP, HCM, SCM, and CX—Oracle enables customers to take advantage of the latest innovations within their existing business processes,” the company concludes.
1 day ago
