Watch a 25-minute technical demonstration that walks through the end-to-end process of fine-tuning a Llama 3.1 model using Weights & Biases (W&B) Models platform. Learn how to leverage W&B Models as an AI system for pre-training, fine-tuning, governing, and managing models throughout their lifecycle from experimentation to production. Follow along with the implementation of experiment tracking, model registry publishing, and automated CI/CD workflows for evaluation and deployment. Explore key concepts including W&B Registry for model and dataset management, experiment tracking capabilities, results visualization in W&B projects, and setting up automated model evaluation using W&B Automations and GitHub Actions. Gain practical insights into streamlining the machine learning workflow to accelerate time to market while maintaining proper governance and management of AI models.
End-to-End LLM Fine-Tuning with Weights and Biases Models - From Experimentation to Production
Overview
Syllabus
Introduction to W&B Models
Reviewing CI/CD for machine learning
Introduction to W&B Registry
Description of the LLM fine-tuning use case
Experiment tracking with W&B Models
Viewing results in a W&B project
Publishing a model to W&B Registry
Using W&B Automations to kick-off CI/CD workflows
Model evaluation using W&B Automations and GitHub Actions
Reviewing model evaluation results using W&B Reports
Model deployment into production and conclusion
Taught by
Weights & Biases
Related Courses
-
Optimizing CI/CD Model Management and Evaluation Workflows with Weights & Biases
-
Fine-tuning LLMs to Reduce Hallucination - Leveraging Out-of-Domain Data
-
Fine-tuning AI Models for Healthcare Policy Analysis with AWS and Snorkel Flow
-
Fine-Tuning and Deploying LLaMA 3.1 on Hugging Face and Ollama
-
Synthetic Data Generation and Fine-tuning for OpenAI GPT-4 or Llama 3
-
Stanford ALPACA 7B LLM: Fine-tuning Guide with Code and Datasets
