#127127th of 206 in ML frameworks and librariesFor data and documents, data analysis
pytorch-forecasting
PyTorch library for time series forecasting with neural network architectures
- Stars
- 5k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Jun 2026
- Language
- Python
4,838 on 22 Apr 20265,011 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Python package that trains deep learning forecasting models on pandas DataFrames through a TimeSeriesDataSet class and PyTorch Lightning. It includes Temporal Fusion Transformer, N-BEATS, N-HiTS, DeepAR, PatchTST, LSTM/GRU and MLP baselines, plus multi-horizon metrics and Optuna hyperparameter tuning. Training runs on CPU or on one or more GPUs, with TensorBoard logging.
Who it is for: ML engineers and researchers building neural time series forecasters
Strengths
- TimeSeriesDataSet handles transformations, missing values and variable history lengths
- Ships TFT, N-BEATS, N-HiTS, DeepAR and PatchTST with interpretation plots
- Lightning-based training on CPU, single or multi-GPU with TensorBoard logging
- Built-in Optuna tuning and a tutorial for implementing custom models
Weaknesses
- Forecasting only; no general-purpose model or LLM support
- Requires PyTorch and Lightning; Windows needs a manual PyTorch install step
- Input must be a pandas DataFrame with integer time index and string series IDs
- Model set is limited to the listed architectures plus custom implementations
What it needs
- GPU optional
- Needs PyTorch, PyTorch Lightning, pandas, Optuna, TensorBoard
- Models: Temporal Fusion Transformer, N-BEATS, N-HiTS, DeepAR, PatchTST
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in ML frameworks and libraries
See all 206| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | ultralyticsPython library and CLI for training and running YOLO vision models | 94 out of 100 |
| 3 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |