README.md

Vision Fine-tuning examples

This asset collection is intended to provide reusable examples for vision fine-tuning based on Azure OpenAI and custom OSS models. We have started by providing an end to end example of GPT-4o fine-tuning based on a public dataset and intend to add further examples in the near future.

Get started

Create and activate a virtual Python environment for running the app. The following example shows how to create a Conda environment named vision-ft:

conda create -n vision-ft python=3.12
conda activate vision-ft

Install the required packages. Navigate to the vision-fine-tuning folder and execute the following:

pip install -r requirements.txt

Required Services:

• An Azure OpenAI resource with the following model deployments:
  • GPT-4o

Optional Services:

• Azure AI Foundry

Navigate to the starter notebook:

• Fine-tuning Azure OpenAI GPT-4o for Chart Analysis