Explore the world of open-source Large Language Models (LLMs) and their practical applications in this comprehensive 1-hour 7-minute webinar. Gain insights into the comparative analysis between open-source LLMs and proprietary options like OpenAI, understand the economic aspects of hosting these models, and learn how to evaluate and benchmark LLM performance. Delve into topics such as code generation, memory calculation, quantization, Llama 2 latency and cost, supervised fine-tuning, reinforcement learning for human feedback, and the differences between RLHF and DPO. Discover the potential of open-source LLMs for production environments and explore tools like dstack to enhance your understanding of this rapidly evolving field.
Leveraging Open-Source LLMs for Production
Overview
Syllabus
Introduction
Why open-source LLMs?
Major open-source LLMs
Code generation
Memory calculator
Quantization
Llama 2 latency and cost
Supervised fine-tuning
Reinforcement learning for human feedback
RLHF vs DPO
dstack
Taught by
Data Science Dojo
