In this video, I show how you can easily migrate your applications from OpenAI models to Hugging Face open-source models.

First, we deploy a Llama3 8B model on an Inference Endpoint hosted on Google Cloud, and we invoke it with the OpenAI client library.

Then, we deploy a Zephyr 7B model on Amazon SageMaker and invoke the endpoint with the OpenAI Messages API.

It couldn’t be simpler to enjoy the many benefits of open-source models, without having to rewrite your app :)