Now that the fine-tuned model is ready, we need to integrate it into the streamlit app.
Set up chatbot UI
Set up a simple user interface for the user to interact with your own fine tuned model. Nothing fancy.
Connect to model API
Your fine-tuned model will be hosted on Open AI API, so you will have to authenticate that it's you making requests to your model. To do this, you will need to:
- Use OpenAI, Hugging Face, or your custom endpoint
- Implement POST requests for sending user messages and receiving model responses
- Handle authentication securely (API keys, env variables)
You have already done something similar with the Open API keys to call the gpt model.
OPTIONAL: Deploy chatbot
You can try deploying the app on streamlit cloud, which has a dedicated for hosting streamlit apps.
For this, you will need to:
- Set up consistent folder structure in the repo for streamlit cloud to access your files
- Add the environment variables to the streamlit server, including API keys and such
Now that the fine-tuned model is ready, we need to integrate it into the streamlit app.
Set up chatbot UI
Set up a simple user interface for the user to interact with your own fine tuned model. Nothing fancy.
Connect to model API
Your fine-tuned model will be hosted on Open AI API, so you will have to authenticate that it's you making requests to your model. To do this, you will need to:
You have already done something similar with the Open API keys to call the gpt model.
OPTIONAL: Deploy chatbot
You can try deploying the app on streamlit cloud, which has a dedicated for hosting streamlit apps.
For this, you will need to: