Lightweight Text Classification using CNN and Sinusoidal Positional Ecoding

This is the replication of the paper Lightweight Text Classifier using Sinusoidal Positional Encoding.

Folders

• 1 - code contains code for individual models and data preparation process. load_data.py contains functions to load datasets. Individual jupyter files are named by model name.
• 2 - result contains original output json file from the training process of each model.
• 3 - tables contains summarized result for each model.