YCWPD 101 - Large Language Models (LLMs) and Prompt Engineering
Description
Explore the transformative power of large language models (LLMs) in this comprehensive workshop designed for business professionals. Discover how these advanced AI tools can address real-world business challenges with efficiency and accuracy. This workshop offers a hands-on approach, guiding participants through practical applications of LLMs to solve complex business problems.
Learn how to harness the potential of tools like Langchain's LLM wrapper and Chainlit's LLM application builder. Engage in interactive exercises, coding tutorials, and insightful discussions that illuminate the practical aspects of LLM technology. Participants will also work with industry-standard frameworks, such as Hugging Face's Transformers library, gaining valuable experience in implementing LLM solutions for business use cases. Whether you are new to AI or looking to deepen your understanding, this workshop provides a thorough and practical guide to leveraging large language models in business settings.
Topics Covered
- Understanding Large Language Models: Develop an understanding of the capabilities of large language models and gain insights into pre-trained models such as LLaMA, GPT, and more.
- Exploring Prompt Engineering: Discover the crucial role of prompts in interacting with large language models and achieving desired outputs. Learn how to design effective prompting techniques that interface with LLMs.
- Exploring LLM Wrappers: Dive into the techniques such as chaining LLMs, agents, and utils using LangChain.
- Hands-On Coding Experience and Model Interaction: Engage in practical activities that allow you to interact with large language models using various frameworks and libraries. Learn how to answer questions, summarize documents, facilitate search engines for online and offline queries, enable SQL queries across different database and file structures, and extract information from text and PDF files using LLM agents with OpenAI and Hugging Face embeddings.
- LLM Applications: Create and Collaborate on Large Language Model Apps. In this context, we delve into the potential of Chainlit for developing LLM applications within the Langchain environment.
- Evaluating Model Outputs: Explore metrics such as perplexity (PPL), BLEU and ROUGE as common metrics for evaluating language models.
Learning Outcomes
Upon completion of this workshop, you should be able to:
- Explain large language models and their pre-training process
- Gain proficiency in prompt engineering techniques
- Develop practical skills in interacting with large language models for various tasks with Langchain and Chainlit.
- Evaluate and assess large language model outputs effectively
Notes
You may bring your own laptop or use one of the computers provided in class.
Participants are solely responsible for acquiring their own tokens to run code during or after the workshop. To access online LLMs, you will need an access token or API key from Hugging Face and OpenAI.
Obtain a free trial token and/or secure payable API by visiting the following links:
Hugging Face: https://huggingface.co/docs/hub/security-tokens