TRIADS Training Series: Large Language Models in Python

In this 3-session seminar, we will delve into neural network/deep learning approaches that have become dominant in NLP in recent years. The course will cover concepts in the rapidly evolving state-of-the-art, with a focus on transfer learning using pretrained language models and transformers such as BERT and GPT. Participants will leverage these models off the shelf to accomplish tasks like translation and zero-shot classification. Additionally, we will explore the functioning of hyperparameters and learn how to train these models for specialized tasks. All sessions will be conducted in Python, and code walkthroughs will be provided to help participants grasp the implementation of these models. Participants are expected to have a basic proficiency in building machine learning models that use text data and some understanding of how neural networks work.  

This class will be fully in-person, and participants will use their own laptops.

TRIADS training workshops are co-sponsored by University Library Data Services, as part of the DataLab Workshops series.

Time: Monday, Tuesday and Wednesday, 12:30 - 2 p.m.

Location: Olin Library Instruction Room 2

Instructor: Ishita Gopal

Max enrollment: Enrollment is limited to 30. If you enroll and elect not to attend, please let us know ASAP so we can offer the space to another participant.