This talk dives into the practical use of open-source LLMs for real-world applications, weighing their pros and cons. It highlights advantages like privacy and cost-efficiency, while also touching on challenges such as the need for technical know-how and techniques to efficiently train and deploy LLMs.
The talk also looks at the economics of fine-tuning and deploying open-source LLMs, covering costs and infrastructure needs. It discusses various frameworks for deploying these models and gives an overview of key open-source LLMs, emphasizing their features and capabilities.
This talk aims to give attendees a general technical understanding of how to effectively leverage open-source LLMs in their projects.
Speaker
Andrey Cheptsov
Founder and CEO @dstack, Rethinking Container Orchestration for AI Workloads, Previously Product Manager @JetBrains Working on PyCharm and DataSpell
Andrey is the core contributor to dstack. He is passionate about open-source and AI. Previously, Andrey worked at JetBrains, incl. the PyCharm team.