This is an exercise to implement a few deep learning models for recognising human activities (like walking, standing, etc) from sensor data (accelerometer and gyroscope) collected from mobile phones.
We use Jason Brownlee's tutorial (in Keras) as a source of inspiration and look to implement the models in PyTorch.
Python 3.7
Pytorch 1.4
pandas
scikit-learn
seaborn
jupyter
tensorflow 2.1 (if you are running the tensorflow notebook)
- pip install the obove required libraries
- Open terminal and cd to this root folder
- Run jupyter notebook from the root folder
- Open the human_activity_recognition.ipynb notebook
- The notebook is fairly self explanatory.