NeMo Customizer RL Training Container

This container provides alignment training functionality for the NeMo Customizer ecosystem. It is specifically designed to run model alignment and customization jobs using Direct Preference Optimization (DPO).

The NeMo Customizer RL image integrates NeMo RL framework with the Customizer training infrastructure, enabling advanced post-training techniques that align language models with human preferences and specific task requirements. It supports distributed training across multiple GPUs and nodes.

Resources

Helm Chart | User Guide

Note: Use, distribution or deployment of this microservice in production requires an NVIDIA AI Enterprise License.

Governing Terms

The software and materials are governed by the NVIDIA Software License Agreement and the Product-Specific Terms for NVIDIA AI Products.