# How to fine-tune Llama models for structured extraction of data from images

Extracting data from images can be a challenge, but with Llama models, it doesn't have to be. In this tutorial, you'll learn how to fine-tune Llama 11B Vision Instruct to improve performance for image-to-JSON data extraction. We will go over how to customize and evaluate Llama models for this specific use case while preserving existing knowledge.

You’ll learn how to:

• Extract structured data from images using vision models  
• Achieve performance improvements with minimal adjustments  
• Prepare and split datasets for model training  
• Evaluate base model performance  
• Understand model training configurations and their components  
• Verify that specialized training doesn't negatively impact general performance

Start customizing Llama for your use case today! Head to the GitHub recipe to get started:

[Explore the GitHub recipe](https://github.com/meta-llama/llama-cookbook/tree/main/getting-started/finetuning/vision)

## Explore more

**How to build your own multi-modal WhatsApp chatbot with Llama 4**  
Build an intelligent chatbot for your business on WhatsApp.  
[Watch now](/content/resources/videos/build-multi-modal-whatsapp-chatbot-with-llama-4/index.html)

**How to quickly generate technical blogs from your docs with Llama and RAG**  
Harness Llama & RAG to automate technical blog writing. Generate accurate, well-structured posts from your existing documentation, saving time & effort.  
[Watch now](/content/resources/videos/generate-technical-blogs-from-your-docs-with-Llama-and-RAG/index.html)

**How to implement Llama Protections to your GenAI applications**  
Safeguard your GenAI apps with Llama Protections. Learn to integrate open-source solutions for content moderation, jailbreak defense, and code security.  
[Watch now](/content/resources/videos/implement-llama-protections-to-genAI-applications/index.html)