This is a GUI wrapper for the HuggingFace Pipeline for zero shot classifiers
This project was designed as a way for me to practice GUI developement, and AI pipelining. This uses the facebook/bart-large-mnli model from huging face as a zeroshot classifier (found in the classifier_class module). The GUI.py file defines the logic to create a graphical user interface for the bart-large-mnli model as a zeroshot classifier. The GUI makes this tech much more accessable to people of all techonological levels.
Overall this project was great practice, and a fun weekend project to teach myself how to use hugging face models programatically, and how to develope simple GUI applications. I recognize that the steps involved to install this application work against the idea of using a GUI to simplify things. I am working on making an executable release for this, so it can be insatlled and run without much effort.
I recomend initilizing a virtual environment for instling this project
In addition to the requierments.txt, you are going to need:
After you have installed these things, you can pip install -r requierments.txt
This should be enough to get the project working. If you are running from a virtual environment, make sure you start GUI.py from an activated terminal.
After installing everything, I recommend you make a batch file that enables the virtual environment, and runs the GUI.py module. Here is an example one liner script:
cmd /c "cd /d %~dp0 & cd /d .\venv\Scripts & activate & cd /d ..\.. & python GUI.py"
If you paste this into a batch file in the same directory as GUI.py and the virtual environment (assuming a correct instal), then this script should start you up without a problem.