1. Home
2. Data Science
3. Time Series Analysis
4. PyTorch Forecasting

PyTorch Forecasting

An implementation on top of PyTorch of several deep learning models for time series forecasting. Models include Temporal Fusion Transformers, N-BEATS, DeepAR, and simple standard networks for baselineing.

Logo
Website	https://pytorch-forecasting.readthedocs.io/en/latest/#
Repository	https://github.com/jdb78/pytorch-forecasting
Byline	Time series forecasting with PyTorch
License	MIT
Project age	2 years 9 months
Backers	Personal Project (Creator and maintainer)
Lastest News (2022-03-23)	v0.10.0 Adding N-HiTS network (N-BEATS successor) more
Size score (1 to 10, higher is better)	3.5
Trend score (1 to 10, higher is better)	6.5

Education Resources

URL	Resource Type	Description
https://towardsdatascience.com/introducing-pytorch-forecasting-64de99b9ef46	Blog	Overview of the PyTorch-Forecasting project by the project creator.

Git Commit Statistics

Statistics computed using Git data through November 30, 2022.

Statistic	Lifetime	Last 12 Months
Commits	18,605	4,534
Lines committed	3,448,602	733,470
Unique committers	35	13
Core committers	6	5

Similar Projects

Project	Size Score	Trend Score	Byline
Darts	5.25	6.25	A python library for easy manipulation and forecasting of time series.
GluonTS	3.5	8.75	Probabistic time series modeling in Python
Kats	4.5	5.25	Kats is a toolkit to analyze time series data, a lightweight, easy-to-use, and generalizable framework to perform time series analysis.
Merlion	2.75	5.5	A Machine Learning Framework for Time Series Intelligence
Orbit	3.25	5.5	A Python package for Bayesian forecasting with object-oriented design and probabilistic models under the hood.