Discover a layer pruning technique for Large Language Models (LLMs) that enhances cost efficiency in domain adaptation in this 17-minute conference talk from MLOps World: Machine Learning in Production. Explore the PruneMe repository, inspired by "The Unreasonable Ineffectiveness of the Deeper Layers," and learn how removing redundant layers facilitates continual pre-training on streamlined models. Understand the process of merging these models into a top-performing general model using advanced techniques like Evolve Merging. Gain insights into this cost-effective approach to model optimization and adaptation, presented by Shamane Siri, Ph.D., Head of Applied NLP Research at Arcee.ai.
Enhance Cost Efficiency in Domain Adaptation with PruneMe
MLOps World: Machine Learning in Production via YouTube
Overview
Syllabus
Enhance Cost Efficiency in Domain Adaptation with PruneMe
Taught by
MLOps World: Machine Learning in Production
Related Courses
-
Large Language Models with Azure
-
H2O ai Large Language Models (LLMs) - Level 2
-
Master AI Efficiency with LoRA - Optimize Fine-Tuning for Large Language Models
-
RAG vs PEFT - Comparing LLM Domain Adaptation Methods
-
Alignment is Real - Fine-Tuning vs. Prompting in Domain-Specific Tasks
-
Fine-Tuning LLMs: Best Practices and When to Go Small - Lecture 124
