Inference as a Service¶

Overview¶

Inference as a Service is designed to simplify the process of utilizing pre-trained models for making predictions on your data. It offers a suite of tools that enable users to explore, select, and deploy models into production with ease.

Components of Inference as a Service¶

Playground¶

• Description: An interactive platform for experimenting with various machine learning models without coding.
• Features:
• Test models on your own data.
• Adjust model parameters via the UI.
• Suitable for non-developers and AI enthusiasts.

Model Market¶

• Description: A curated collection of pre-trained models available for private use.
• Features:
• Browse and select models based on specific needs.
• Deploy models on new resources with ease.
• Models optimized for tasks like NLP, complex reasoning, and conversational applications.

Model Service¶

• Description: A tool for deploying machine learning models as API services.
• Features:
• Deploy public and private models.
• Easy deployment, API access, auto-scaling, and performance monitoring.
• Deploy models as APIs for real-time inference.
• Scale inference with GPU-backed infrastructure.
• View usage metrics for reliability and efficiency.

Next Steps¶

To learn more about each component and how to utilize them for your AI development needs, explore the detailed guides for: - Playground - Model Market - Model Service