OCI Generative AI - Overview

Oracle Cloud Infrastructure Generative AI (Beta) is a fully managed service that provides a set of state-of-the-art, customizable large language models (LLMs) that cover a wide range of use cases for text generation.

“Oracle is launching a generative AI cloud service for enterprise customers. This new service protects the privacy of our enterprise customers' training data, enabling those customers to safely use their own private data to train their own private specialized LLM.

The OCI Generative AI service includes the following foundational models:

Generation

Give instructions to generate text or extract information from your text.

Summarization

Summarize text with your instructed format, length, and tone.

Embedding

Convert text to vector embeddings to use in applications for semantic searches, text classification, or text clustering.

Using the Pretrained Models

OCI Generative AI provides "Playground" to try the pretrained models. You can explore the Playground running your use cases with your Prompts, adjust the parameters, update your prompts, and rerun the models, without taking care about the infra. Once you are satisfied with the result, you can then get the code from the Console and copy the code into your applications.

Fine-Tuning the Pretrained Models

Once you have the code of a pre-trained model and train the model with your own training dataset (uploaded on OCI Object Storage) and leverage OCI GenAI powerful Dedicated AI Clusters specially sized for fine-tuning. Here, your model will run on Dedicated AI Cluster hosted on your tenancy, completely satisfying the need of enterprise industry.

This fine-tuned model can be integrated smoothly with enterprise application by creating an endpoint for these custom models.

Key Highlights for Enterprise Customers:

• Customer data privacy is respected. When you send data to the Generative AI API, it’s processed in memory and no customer data is stored. Oracle temporarily logs some metadata about your API requests to improve the service, for billing and metering, and to combat abuse. The metadata is, for example, the time and size of the request.
• Customer data privacy is respected. No other customer can look at your data or models.
• Local file storage for text embeddings and Oracle Object Storage for fine-tuning custom models are supported.

Note: Currently, OCI Generative AI service is in pre-GA (Beta) releases and it's free to explore until it's GA.