TWYM is a final project for Neural Networks and Deep Learning course.
It is based on data from Sussex-Huawei Locomotion Challenge.
In this contest, participants try to predict one of eight locomotion modes (Still, Walk, Run, Bike, Car, Bus, Train, Subway) based on readings from various sensors, such as accelerometer or gyroscope.
We wanted to compare state-of-the-art IndRNN model of one of the winners with approach in which we classified time series based on their plots with Siamese CNN.
The repository contains a report describing our approach and results.