Large Language Model Fine-Tuning using Parameter Efficient Fine-Tuning (PEFT)

We all recognize the immense business opportunity from generative AI and large language models (LLMs) — particularly those trained or developed on proprietary company data. However, developing them is resource-intensive, time-consuming, and requires deep technical expertise. The NVIDIA NeMo Framework accelerates LLM training by up to 30% (models ranging from 22 billion to 1 trillion parameters) — a quick, efficient, containerized framework for model training, evaluation, and inference. In this session, we'll use NVIDIA NeMo Megatron — a powerful transformer developed by the Applied Deep Learning Research team at NVIDIA — to fine-tune using parameter efficient fine tuning (PEFT) on Domino’s Enterprise AI Platform. We’ll walk through the end-to-end model life cycle, starting with NVIDIA’s NeMo Toolkit in Domino’s AI Project Hub, customizing a data science environment for fine-tuning in Domino, then use NeMo Megatron to encode text prompts and generate task-specific virtual tokens.