Skip to main content
Campus Calendar

Large Language Model Training and Development

  • To
  • To
  • Van Munching Hall
LLM Training & Development

Course Overview

Join Professor KZ Zhang for a deep dive into the technical framework behind the large language model (LLM), including solutions for developing and training a model on a custom data set. The solution can be implemented in any environment, even without extensive computing resources.

Who should take this course?

While this course is intended for anyone interested in how an LLM works, it is especially useful for industries with restrictive data privacy rules, such as government, federal contracting, or healthcare. Participants should be familiar with Python and linear algebra (e.g. matrix multiplication). Participants will be asked to apply lessons learned in real-time coding exercises.

Why Build Your Own Model?

Industry-Specific Accuracy: Develop AI models that understand complex terminology and regulations unique to your industry, ensuring precise and reliable results.

Data Security & Compliance: Keep sensitive information confidential and compliant by training models on internal data.

Operational Efficiency: Streamline processes and enhance decision-making with custom AI solutions designed for the specific needs of your environment.

Valuable Outcomes

Define the foundations of LLMs/transformers.

Apply a pre-trained and fine-tuned paradigm in text understanding.

Deploy a customized LLM for your organization.

Course Dates

Day 1: Friday, August 2, 2024, 9:00 AM - 2:30PM

Day 2: Friday, August 9, 2024, 9:00 AM - 2:30PM

Day 3: Friday, August 16, 2024, 9:00 AM - 2:30PM

Day 4: Friday, August 23, 2024, 9:00 AM - 2:30PM

Daily Agenda

9:00 AM Workshop begins

12:00 PM Lunch break (Boxed lunch provided)

1:00 PM Workshop resumes

2:30 PM End of day

Location

Van Munching Hall

Contact

Office of Executive Education (Smith School of Business)

For disability accommodations, please contact Clayton Richey at crichey@umd.edu

Event Tags

Schools and Units

Audience

Tags

Event Topics