Metraipolis › archive › dc_c032d520b7fa

Tabby: An Open Pretraining Recipe for Time Series Foundation Models — arXiv:2609.13956

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.13956
License (as archived): CC0-1.0
Archived: 2026-10-16T00:10:35.927Z
SHA-256 of the archived text: 094921c92385390b3de7b3b6209a5c82e933578f3fb3662e3798fd85ff331df3
Stable URL of this page: https://metraipolis.com/documento/dc_c032d520b7fa

How to cite

Tabby: An Open Pretraining Recipe for Time Series Foundation Models — arXiv:2609.13956 — arXiv (en), https://arxiv.org/abs/2609.13956 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c032d520b7fa (Metraipolis archive dc_c032d520b7fa, sha256 094921c92385390b3de7b3b6209a5c82e933578f3fb3662e3798fd85ff331df3)

BibTeX

@misc{dcc032d520b7fa,
  title = {Tabby: An Open Pretraining Recipe for Time Series Foundation Models — arXiv:2609.13956},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.13956},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c032d520b7fa; sha256 094921c92385390b3de7b3b6209a5c82e933578f3fb3662e3798fd85ff331df3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.13956. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In this report, we release Tabby, a long context probabilistic time series foundation model, together with a complete and open recipe of how it was built. Tabby adopts an encoder-only patch Transformer architecture and concentrates the contributions on the data and the training procedure. The pretraining corpus combines an extended real-world collection, GIFT-Eval-Pretrain+ and BLAST, with synthetic data from KernelSynth and CauKerV2, an online generator that composes temporal dynamics through randomly sampled structural causal models. Training couples a progressive convergence schedule, which yields reusable intermediate checkpoints, with a deep quantile supervision objective for intermediate layers. The resulting 145M parameter backbone supports contexts of up to 8,192 observations and serves forecasting, classification, and anomaly detection, while a prompt-tuning module further improves in-distribution forecasting performance with the pretrained weights frozen. Tabby achieves competitive zero-shot forecasting performance on GIFT-Eval and the out-of-distribution TIME benchmark, while the same pretrained backbone also supports classification on the UCR Archive and zero-shot anomaly detection on TSB-AD-U. We release training pipeline and model as open source at huawei-noah/trustworthyAI.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.13956 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T23:54:14.667Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).