Recurrent Domain Incremental Adaptation of Time-Series Foundation Models
Paper • 2611.24075 • Published • 3
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Processed train and test splits used to reproduce the PACT results in arXiv:2611.24075.
Project page: https://sachithx.github.io/research/pact.html
Code: https://github.com/Sachithx/PACT. The PACT repository's download_data.py pulls this dataset into dataset/.
The frozen backbone checkpoint is thuml/sundial-base-128m. It is downloaded by the training script, not stored here.
dataset/FD/train_{0..3}.pt dataset/FD/test_{0..3}.pt
dataset/WISDM/train_{0..35}.pt dataset/WISDM/test_{0..35}.pt
dataset/HHAR/train_{0..8}.pt dataset/HHAR/test_{0..8}.pt
dataset/UCIHAR/train_{1..30}.pt dataset/UCIHAR/test_{1..30}.pt
dataset/Sleep-EDF/train_{0..19}.pt dataset/Sleep-EDF/test_{0..19}.pt
Each .pt file is a dict with samples and labels.
@misc{abeywickrama2026recurrentdomainincrementaladaptation,
title={Recurrent Domain Incremental Adaptation of Time-Series Foundation Models},
author={Sachith Abeywickrama and Peiliang Gong and Emadeldeen Eldele and Min Wu and Chau Yuen},
year={2026},
eprint={2611.24075},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2611.24075}
}