Understand the project
Practical overview
When it fits
Use it when you need a code implementation to study or adapt for long-sequence forecasting experiments, rather than a hosted forecasting service.
Who it is for
- Researchers reproducing or adapting Informer forecasting experiments.
- Developers evaluating Transformer-based methods on sequential datasets.
Documented uses
- Train and test Informer on the documented ETTh1, ETTh2 and ETTm1 datasets.
- Explore prediction and visualization workflows through the linked Colab examples.
- Adapt the implementation to experiment with domain-specific time-series data.
Before you choose
Requirements & limits
- The README specifies Python 3.6 and PyTorch 1.8.0, alongside pinned versions of matplotlib, NumPy, pandas and scikit-learn.
- The README lists cuDNN 7.3.1.
- The repository identifies Apache-2.0, while the README badge states CC BY-NC-SA; review the licensing materials for the intended use.
- Dataset inputs are zero-mean normalized in this implementation.
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
Run a documented ETT training and testing command
Use the repository's stated ETTh1 command to train and test Informer with ProbSparse attention.
Before you start
- Python 3.6
- Dependencies listed in requirements.txt, including PyTorch 1.8.0
- ETTh1 data files prepared in the repository's expected data location
Steps
Install the listed dependencies with the command below.
pip install -r requirements.txtPrepare the ETTh1 dataset files; the README describes obtaining ETT data from the ETDataset repository and placing required files in
data/ETT/.Run the command below.
python -u main_informer.py --model informer --data ETTh1 --attn prob --freq h
Expected outcome
The documented command starts an Informer training and testing experiment on ETTh1 using ProbSparse self-attention.
Guide limitations
- The README notes that dataset inputs are zero-mean normalized.
- The guide reflects repository instructions; it does not verify successful execution or results.
Evidence and transparency
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 15 October 2026
- Original website title
- zhouhaoyi/Informer2020 · Official homepage title · retained in its source language