/webUI-for-LLM

Simple web UI for API-based LLM inference for models available on HuggingFace

Primary LanguagePython

webUI-for-LLM

The purpose here is to build a simple web UI for conversing with LLM of your choice. This is the first step to build a more sophisticated chatbot, and the aim here it to explore and understand the capabilities of models that have their APIs available for free.

To run this, update the authorisation_file with your access token and the API URL from Hugging Face.

Install the requirements by running the following on your terminal:

pip install -r requirements.txt

Run the main.py script to launch the web UI locally:

python main.py

The quality of conversation depends on the model being used for inference. Smaller models can be loaded quickly but the responses tend to be incoherent. Large models like flan-t5-xxl are too large to load.