All Models

The NVIDIA NeMo™ Framework has a suite of model agnostic tooling that will aid in the following:

• Preparing your environment to get started with the NVIDIA NeMo™ Framework

• Preparing your data

• Logging

• Using the AutoConfigurator

• Resuming training

• PEFT

• Exporting models

Each of the tasks is outlined in detail in the following sections.

• Prepare Environment
• Data Preparation
• Monitoring and Logging
• AutoConfigurator
• Resuming Training with a Different Number of Nodes
• Generalized PEFT Framework
• Exporting NeMo Models to TensorRT-LLM