Building a Chat App with OpenAI, Comet ML, and Streamlit

Author(s): Evren Ozkip

Originally published on Towards AI.

Learn how to create a blog title generator app.(Image by Author)

Training large language models from scratch is difficult because it requires massive amounts of data, time, and powerful computers. But with the help of APIs, you can use pre-trained language models.

Today, we’ll build a blog title generator app with the OpenAI API and Streamlit. We’ll also monitor our LLM experimentation with CometLLM.

The app we’re going to create looks like this:

The app we’re going to build

Here are the topics we’re going to cover in this article:

SetupBuild the backend with OpenAI and CometLLMCreate the frontend with StreamlitTest the app

Let’s dive in!

Before building the app, we need to install the tools we’ll use.

!pip install openai streamlit comet_llm python-dotenv

First of all, we’re going to set the API keys as environment variables with the dotenv library. To build the app, we’re going to leverage the OpenAI API key and CometML API key.

With the OpenAI API key, we’re going to access the GPT model. To get this API key, you can go to the OpenAI website and generate a key.

To log our LLM experimentation, we’re going to use CometLLM. CometLLM is an LLMOps tool that allows you to track your LLM prompts.

If you don’t have a Comet account yet, create one for free here.

After… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI