This repository contains the implementation of a neural networks and deep learning assignment in PyTorch. The assignment is divided into two parts:
In Part A, the task is to perform polarity detection from voice recordings. The dataset used for this task is sourced from the National Speech Corpus. Polarity detection involves classifying voice recordings into positive or negative categories.
The PyTorch framework is employed to build and train neural network models for the classification task. The models are designed to process voice recordings and make predictions on the polarity of the spoken content.
The National Speech Corpus dataset is utilized for training and evaluation. It includes a diverse set of voice recordings, providing a comprehensive range of speech patterns and content.
In Part B, the assignment involves a regression problem focused on predicting HDB (Housing and Development Board) flat prices in Singapore. This task requires building a regression model to estimate the prices of residential properties based on various features.
The regression problem is addressed using PyTorch to construct and train neural networks suitable for predicting HDB flat prices. The models are trained on a dataset that includes relevant features such as location, size, and amenities.
The dataset for this part includes information about HDB flats in Singapore, with details on their locations, sizes, and other relevant attributes.