Pytorch Forecasting aims to ease timeseries forecasting with neural networks for both real-world cases and research alike. Specifically, the package provides:

  • A timeseries dataset class which abstracts handling variable transformations, missing values, randomized subsampling, multiple history lengths, etc.
  • A base model class which provides basic training of timeseries models along with logging in tensorboard and generic visualizations such actual vs predictions and dependency plots
  • Multiple neural network architectures for timeseries forecasting that have been enhanced for real-world deployment and come with in-built interpretation capabilities
  • Multi-horizon timeseries metrics
  • Ranger optimizer for faster model training
  • Hyperparameter tuning with optuna

Don't forget to tag @jdb78 in your comment, otherwise they may not be notified.

Authors community post
Data scientist and consultant who loves to keep up with the latest academic research and drive real business impact
Share this project
Similar projects
Time Series Prediction with LSTM Using PyTorch
Time series applied to forecasting on the Airplane Passengers Dataset.
How to Train Your Neural Net
Deep learning for various tasks in the domains of Computer Vision, Natural Language Processing, Time Series Forecasting using PyTorch 1.0+.
Differentiable controlled differential equation solvers for PyTorch with GPU support and memory-efficient adjoint backpropagation.
Temporal Convolutional Networks for Time-Series
We introduce several novels using TCN, including improving traffic prediction, sound event localization & detection, and probabilistic forecasting.
Top collections