OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Neuralforecast

Neuralforecast is an open-source Python library that provides neural network models for fitting time-series data and generating forecasts.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Neuralforecast — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Neuralforecast when a Python forecasting workflow calls for neural models and a fit/predict interface, especially when you want to include covariates, explore probabilistic forecasts or tune models. Its tutorials also cover hierarchical forecasting and extending the library with a model.

Who it is for

  • Python practitioners building time-series forecasting workflows.
  • Analysts comparing neural forecasting models with baseline methods.
  • Teams forecasting series with external covariates or grouped hierarchical structure.

Documented uses

  • Train a neural model on historical observations and generate forecasts.
  • Compare architectures such as NBEATS, LSTM and transformer models.
  • Use static or temporal covariates, including signals such as weather or prices.
  • Produce probabilistic forecasts using quantile losses or parametric distributions.
  • Tune model hyperparameters with the documented Ray or Optuna integrations.
Before you choose

Requirements & limits

The public repository identifies the project as open source under the Apache-2.0 license. The README documents installation through pip or conda.

  • The package offers more than 30 models, ranging from recurrent networks to transformers.
  • It supports exogenous variables and static covariates, as well as interpretability methods for trend, seasonality and exogenous components.
  • Probabilistic forecasting is supported through quantile-loss and parametric-distribution adapters.
  • Automatic model selection and distributed hyperparameter tuning are documented, with integrations for Ray and Optuna.
  • The package is installed into a Python environment; the documented quick start uses a dataframe and explicit model and horizon settings.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Neuralforecast and fit a sample forecast

Install the Python package and use its documented AirPassengers example to fit an NBEATS model and request predictions.

Before you start

  • A Python environment with pip or conda available.
  • The AirPassengers dataframe provided by neuralforecast.utils.

Steps

  1. Install the package with the command below or conda install -c conda-forge neuralforecast.

    Documented command
    pip install neuralforecast

  2. Import NeuralForecast, NBEATS and AirPassengersDF from the modules shown in the README.

  3. Create a NeuralForecast instance with NBEATS, specifying input_size=24, h=12, max_steps=100 and freq='ME'.

  4. Call fit(df=AirPassengersDF) and then predict().

Expected outcome

The example fits an NBEATS forecasting model to AirPassengersDF and requests its predictions.

Guide limitations

  • The README example uses a bundled sample dataframe; it does not document preparation steps for a reader's own dataset.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
Nixtla/neuralforecast · Official homepage title · retained in its source language
How we review sourcesSuggest a correction