https://www.youtube.com/watch?v=L-01TIbkgkE
Parth Doshi (parthsdoshi), Max Lin (maxlincode)
Tag files with multiple classes to easily index new downloads. User can train the program as it indexes their files.
- First, download the Bert pretrained model from https://storage.googleapis.com/bert_models/2018_10_18/uncased_L-12_H-768_A-12.zip
- Place the unzipped pretrained model in the
nn/bert/pretrained/folder. - Navigate to the root of this repo and run
pip install -r requirements.txt. - Run
cd nn/bert/ && python3 bert_server.py. - This will load the Bert model and creates a REST API around it to respond to external requests.
- Next, edit
app/server.pyand change the value ofBERT_SERVERto the appropriate ip address and port. - Finally, navigate to
./appand runpython3 main.py. This will launch the frontend. - Download the chrome extension under
./chrome_extension/srcin to chrome which will allow us to automatically run downloaded documents through the app and tag them.
- Open the Extension Management page by navigating to chrome://extensions.
- Enable Developer Mode by clicking the toggle switch next to Developer mode.
- Click the LOAD UNPACKED button and select the extension directory.