/ML-NLP-BERT-QnA-App

Question and answer app using pre-trained BERT model.

Primary LanguagePythonGNU Affero General Public License v3.0AGPL-3.0

Q&A App

A question and answer app built on a BERT model pre-trained from SQuAD dataset.

How to Run the App

  1. Open a terminal
cd <working_directory>
  1. Clone this repository
git clone https://github.com/leeping-ng/ML-NLP-BERT-QnA-App.git
  1. Download pre-trained model from https://www.dropbox.com/s/8jnulb2l4v7ikir/model.zip and extract to <working_directory>. For example, the path of bert_config.json should be <working_directory>/model/bert_config.json
  2. Ensure that your environment contains the packages listed in requirements.txt
  3. Run the app with the following command
streamlit run app.py

App Interface

The following screenshot from the app shows a Q&A about a wiki page for the movie Pulp Fiction.