It is a prediction model which predicts emotions reflected through speech. It captures audio using the microphone and converts it to text using speech recognition, and then uses a BERT-based sentiment analysis model to predict the emotions associated with the recognized text.
Techstack used in this project: Python libraries- transformers, TensorFlow, speech_recognition
Instruction for installation:
pip install -r requirements.txt
Dataset - A predefined dataset is used in this project which is taken from 'The Hugging man face'.
Since it is a prediction model so it will give the probability of different emotions expressed through a human voice. Output of this project looks like -
In future, I am going to create a web based application for this project and add on some amazing and creative features to it.