Models

Baseline models

`nazuna.models.simple_average.SimpleAverage`

Bases: BaseSimpleAverage

`nazuna.models.simple_average.SimpleAverageVariableDecay`

Bases: BaseSimpleAverage

`nazuna.models.simple_average.SimpleAverageVariableDecayChannelwise`

Bases: BaseSimpleAverage

Autoformer

`nazuna.models.autoformer.Autoformer`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Haixu Wu, Jiehui Xu, Jianmin Wang, and Mingsheng Long. "Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting." In Advances in Neural Information Processing Systems (NeurIPS 2021), vol. 34, 2021. Paper | arXiv | GitHub

DLinear

`nazuna.models.dlinear.DLinear`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Ailing Zeng, Muxi Chen, Lei Zhang, and Qiang Xu. "Are Transformers Effective for Time Series Forecasting?" In Proceedings of the AAAI Conference on Artificial Intelligence (AAAI 2023), vol. 37, pp. 11121-11128, 2023. arXiv | GitHub

`nazuna.models.dlinear.DLinearChannelwise`

Bases: BasicBaseModel

`nazuna.models.dlinear.NLinear`

Bases: BasicBaseModel

`nazuna.models.dlinear.NLinearChannelwise`

Bases: BasicBaseModel

PatchTST

`nazuna.models.patchtst.PatchTST`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Yuqi Nie, Nam H. Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. "A Time Series is Worth 64 Words: Long-term Forecasting with Transformers." In International Conference on Learning Representations (ICLR), 2023. Paper | arXiv | GitHub

iTransformer

`nazuna.models.itransformer.iTransformer`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, and Mingsheng Long. "iTransformer: Inverted Transformers Are Effective for Time Series Forecasting." In Proceedings of the 12th International Conference on Learning Representations (ICLR 2024), 2024. Paper | arXiv | GitHub

Gateformer

`nazuna.models.gateformer.Gateformer`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Yu-Hsiang Chen, Hsiao-Hua Chang, Chia-Wen Chen, Si-An Chen, Hsiang-Fu Yu, and Cho-Jui Hsieh. "Gateformer: Advancing Multivariate Time Series Forecasting through Temporal and Variate-Wise Attention with Gated Representations." arXiv preprint, 2025. arXiv | GitHub

UniTSTLike

`nazuna.models.unitst.UniTSTLike`

Bases: BasicBaseModel

Original Research

This model is based on the following research:

Juncheng Liu, Chenghao Liu, Gerald Woo, Yiwei Wang, Bryan Hooi, Caiming Xiong, and Doyen Sahoo. "UniTST: Effectively Modeling Inter-Series and Intra-Series Dependencies for Multivariate Time Series Forecasting." Transactions on Machine Learning Research (TMLR), 2025. Paper | arXiv

The official source code was not publicly available at the time of writing, so this implementation follows the description in the paper. Activation function and dropout placement are not specified in the paper and follow choices common in PatchTST-style models.