/stepcounting

Primary LanguagePython

Step Counting with Attention-based LSTM

Implementation of Step Counting with Attention-based LSTM

Requirments

  • torch
  • numpy
  • sklearn
  • tqdm

Instructions for Writing the Code:

Ensure you have the file WeAllWalk.pt in your current directory, containing acceleration data from sighted walkers extracted from the WeAllWalk dataset, comprising three components: acceleration signals, lengths of the signals, and number of steps in the signals.

Run the Python script named main.py. Within this script: a. WeAllWalk.pt is loaded b. The functions available in utils.py are used to pre-process the loaded data. c. The model in the class provided in models.py is trained and evaluated using five-fold cross-validation. d. The evaluation results are outputted by implementing the evaluation metrics from metrics.py.