Nixtla/neuralforecast
A Python library providing PyTorch-based neural network models for time-series forecasting.
Not currently ranked — collecting fresh signals.
star history
NeuralForecast offers a comprehensive set of neural forecasting models ranging from classic RNNs to modern Transformers including NBEATS, NHiTS, DeepAR, and temporal fusion transformers. Built on PyTorch, it provides scalable and user-friendly implementations for training and inference with probabilistic forecasting capabilities. The library focuses on performance and robustness for univariate and hierarchical time series prediction tasks.
Frequently asked
- What is Nixtla/neuralforecast?
- A Python library providing PyTorch-based neural network models for time-series forecasting.
- Is neuralforecast open source?
- Yes — Nixtla/neuralforecast is open source, released under the Apache-2.0 license.
- What language is neuralforecast written in?
- Nixtla/neuralforecast is primarily written in Python.
- How popular is neuralforecast?
- Nixtla/neuralforecast has 4.2k stars on GitHub.
- Where can I find neuralforecast?
- Nixtla/neuralforecast is on GitHub at https://github.com/Nixtla/neuralforecast.