Essential Question:
- How can we create customized Language Models (LLMs) using Python to generate text for various applications?
Learning Objectives:
- Explain the importance of language models in natural language processing and text generation
- Experiment with different training datasets to create specialized LLMs for specific tasks
- Assess the quality and coherence of text generated by the customized Language Model
Standards:
- NYS Next Generation Learning Standards RST.2.11-12 — Determine the key ideas or conclusions of a source; summarize complex concepts, processes, or information presented in a source by paraphrasing in precise and accurate terms.
- New York State Learning Standards CDOS 3a — Students will demonstrate mastery of the foundation skills and competencies essential for success in the workplace.
Materials:
Scaffolds:
Bridging Learning Gaps:
- Scholars are given resources to answer questions in lab in various mediums
Extensions:
- Scholars who complete the lab work on their 2nd blog post about debugging challenges in python
Opening Task (30 Minutes)
- Scholars will complete opening task on Schoology covering key terms realating to large language models
- Randomly selected scholar will facilitate review with peers
System Props Review (05 Minutes)
Review from previous day about customizing LLMs using system prompt
LLM Customization Lab (35 Minutes)
Scholar will work on guided lab activity to explore customizing pre existing large language models using python language tools such as:
- LangChain
- Transformers
Guided activity includes practical code examples + guided questions to check for understanding.
Scholars will analyze the given code, and customize it to a self-made scenario to create a custom AI.
Break (10 Minutes)
LLM Customization Lab (60 Minutes)
Scholars continue work on lab, and submit when completed to schoology