In the previous article Part 3, I have shown deploy the fine-tuned model on Azure Kubernetes Service with the Kaito add-on. In this article, I will show manual evaluation with a series of prompts taken from the fine-tuning dataset. This blog post is part of a series.Part 1: Intro and overview of the KAITO fine-tuning …
Continue reading Deep Dive Into Fine-Tuning An LM Using KAITO on AKS – Part 4: Evaluation
Now that I have fine-tuned a model in Part 2, next is to deploy the fine tuned model into a new Kaito workspace. This blog post is part of a series.Part 1: Intro and overview of the KAITO fine-tuning workspace yamlPart 2: Executing the Training Kubernetes Training JobPart 3: Deploying the Fine-Tuned ModelPart 4: Evaluating …
I will continue from the Part 1 to execute the deployment of the fine-tuning workspace job. This blog post is part of a series.Part 1: Intro and overview of the KAITO fine-tuning workspace yamlPart 2: Executing the Training Kubernetes Training JobPart 3: Deploying the Fine-Tuned ModelPart 4: Evaluating the Fine-Tuned Model Let' start the fine …
Continue reading Deep Dive Into Fine-Tuning An LM Using KAITO on AKS – Part 2: Execution
I uncover the process and workings of fine-tuning a large language model on a dataset. Scenario for fine-tuning a model can be on corporate data such as a knowledge base, product information, human resources. I walk you through the process and share my detailed observations. What is Fine-tuning a large language model? Involves adjusting a …
Continue reading Deep Dive Into Fine-Tuning An LM Using KAITO on AKS – Part 1: Intro
I have been exploring some in depth tutorials in building a chat application. It implements Retrieval Augmentation Generation (RAG) on a product database in Azure AI Search. The scenario is for a retail customer to ask product recommendations on camping gear. This solution employs Azure AI Foundry for LLM models and interfacing. My goal of …
Continue reading Building a Chat App using Azure AI Foundry SDK and AI Search
The Azure AI Foundry portal is a cohesive environment for developing AI solutions, enabling users to create Hubs and Projects that facilitate collaboration among data scientists, developers, and engineers. It integrates services like Azure Open AI, offering shared resources and role-based access management while streamlining AI services' deployment and management processes.
The blog discusses the Kubernetes AI Toolchain Operator (KAITO) setup on AKS, detailing the NVIDIA GPU VM instance types for node pools hosting AI inference models. It emphasizes cost efficiency and GPU specifications for deploying large models.
KAITO simplifies the deployment of large language models (LLMs) in Azure Kubernetes Service (AKS) environments with preset GPU configurations. This tool automates the setup process, including node provisioning and identity management, essential for data experiments while ensuring security compliance. It enhances efficiency, allowing engineers to focus on AI/ML model experimentation. #azure #kubernetes #AI #genAI #mvpbuzz
I did an experiment of running a load test against 4 applications and while executing a Azure Kubernetes Cluster Upgrade to see the affects and any downtime. The applications hosted in my AKS cluster are AKS Helloworld - super simple application Voting App - simple application with Redis backend Bookinfo Istio demo - microservices architecture …
If you ever encounter the following error, I'll show you one way that helped me resolve my AKS cluster. The cluster is in a failed state. If you didn't do an operation, AKS may resolve the provisioning status automatically if your cluster applications continue to run. To further troubleshoot, you can refer to https://github.com/MicrosoftDocs/SupportArticles-docs/blob/main/support/azure/azure-kubernetes/cluster-node-virtual-machine-failed-state.md For …
Roy Kim on Azure and AI 