Code patterns similar to Informer et al.

This repository supports the following models for ETTm2:

NBEATS: https://github.com/Nixtla/neuralforecast

NHITS: https://github.com/Nixtla/neuralforecast

DLinear: https://github.com/cure-lab/LTSF-Linear

NLinear: https://github.com/cure-lab/LTSF-Linear

TiDE: https://github.com/google-research/google-research

FiLM: https://github.com/DAMO-DI-ML/NeurIPS2022-FiLM

SpaceTime: https://github.com/HazyResearch/spacetime

MultiResolutionDDPM: https://github.com/dlgudwn1219/mrDiff

Informer: https://github.com/MAZiqing/FEDformer

Autoformer: https://github.com/MAZiqing/FEDformer

FEDformer: https://github.com/MAZiqing/FEDformer

PatchTST: https://github.com/yuqinie98/PatchTST

Pyraformer: https://github.com/ant-research/Pyraformer

Triformer: https://github.com/razvanc92/triformer

Download the model zoo from: https://bit.ly/LHFModelZoo

Use this to run the code:

python run.py --root_path [ETT-small DIR PATH] --data_path ETTm2.csv --model [MODEL] --data ETTm2 --features [S,SM,M] --is_training 0 --pred_len [96,192,336,720] --enc_in [1,7] --dec_in [1,7] --c_out [1,7] --itr [N] --model_params_json trained_models.json

If you found this repository useful, consider citing:

https://arxiv.org/abs/2506.12809: A Review of the Long Horizon Forecasting Problem in Time Series Analysis

https://arxiv.org/abs/2601.02094: Horizon Activation Mapping for Neural Networks in Time Series Forecasting

@article{krupakar2026horizon,
  title={Horizon Activation Mapping for Neural Networks in Time Series Forecasting},
  author={Krupakar, Hans and Kandappan, VA},
  journal={arXiv preprint arXiv:2601.02094},
  year={2026}
}

@misc{krupakar2025reviewlonghorizonforecasting,
      title={A Review of the Long Horizon Forecasting Problem in Time Series Analysis}, 
      author={Hans Krupakar and Kandappan V A},
      year={2025},
      eprint={2506.12809},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2506.12809}, 
}