Mainly, this project shows how to build a text classification using pretrained model, which is DistilBERT, and finetuning it using desirable data to output desirable output as well.
Dataset:
Pretrained Model:
- DistilBERT
Finetuned Model:
But, as a product of this project, I have built a chrome extension. Below are steps how to run the program (product):
- go to 'chrome-extension'
- you will find main.py, which is the server (using flask), and run it (but don't forget to activate the conda environment which all necessary packages can be found inside a yml file)
- to make the chrome-extension to appear, you need to go to 'extension' menu in chrome setting, and activate the developer mode
- once activated, you will notice on top left, there will be a menu asking you to upload unpacked chrome-extension
- upload the 'chrome-extension', and turn it on
- once the extension is on, try to select any text on any website (don't forget to refresh the page if it's already opened before the extension)
- then, once the mouseup, there will be a popup appears right under your cursor that tells the emotion of the selected text