Question-Answering-Bert

Demonstrating video

The app takes two inputs:

1. Speech or text question.
2. Text to search for the answer in it.

Used Modules:

• Answer questions: BERT fine-tuned on The Stanford Question Answering Dataset (SQuAD).
• Speech to Text: Google Speech Recognition via SpeechRecognition library.
• Recording: sounddevice library.
• GUI: Tkinter.