The problem Emotion Classifier solves

This project is a Speech Emotion Classifier built using Machine Learning. It Classifies 6 categories of speech using 2 different models. We are extracting certain features like MFCC, Mel spectrogram, chroma_stft, and chrom_cqt, to predict the emotions. We are using libraries like Librosa, Sklearn, Pandas, and Numpy to process data. The model is then deployed on the backend using Streamlit.

Challenges we ran into

While developing the project we ran into certain difficult situations: It was quite difficult to work on audio data, moreover since the Hindi language was given the priority it was a challenge to find a suitable dataset for the model. We ran into problems while writing the backend of the project. But eventually, we completed it.