Metraipolis › archive › dc_0d2fbda37426
NanoForecast v0.5: Competitive Time Series Forecasting Through Training Pipeline Optimization — arXiv:2609.31669
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.31669
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T22:36:04.713Z
- SHA-256 of the archived text
3d2b4908c356a9a1a273fd3cef65ff67ae23179cb4124ea6538e3e73c151c06f- Stable URL of this page
https://metraipolis.com/documento/dc_0d2fbda37426
How to cite
NanoForecast v0.5: Competitive Time Series Forecasting Through Training Pipeline Optimization — arXiv:2609.31669 — arXiv (en), https://arxiv.org/abs/2609.31669 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0d2fbda37426 (Metraipolis archive dc_0d2fbda37426, sha256 3d2b4908c356a9a1a273fd3cef65ff67ae23179cb4124ea6538e3e73c151c06f)
BibTeX
@misc{dc0d2fbda37426,
title = {NanoForecast v0.5: Competitive Time Series Forecasting Through Training Pipeline Optimization — arXiv:2609.31669},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.31669},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0d2fbda37426; sha256 3d2b4908c356a9a1a273fd3cef65ff67ae23179cb4124ea6538e3e73c151c06f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31669. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We present NanoForecast v0.5, a 6.5M-parameter forecaster that competes with models 31x its size (TimesFM, 200M parameters) after training pipeline fixes and no architecture change. Retraining the v0.3 architecture with corrected loss-scope handling, tensor shape alignment, and wider augmentation coverage cuts overall Mean Absolute Scaled Error by 43.8% under one fixed protocol (MASE 3.030 to 1.704) on the same data and compute budget. NanoForecast v0.5 beats TimesFM on all three ETT datasets (MASE 0.676/1.110/0.287 vs. 0.705/1.360/0.545) and on exchange rate (4.317 vs. 4.383); TimesFM keeps a clear lead on the high-cardinality electricity and traffic sets. Against PatchTST (15M+ parameters, official configuration), v0.5 wins all three ETT sets. Training takes about 12 hours on a single cloud GPU (NVIDIA T4, Google Colab) and inference needs no GPU (measurements in this paper are on an Apple M4 CPU). We release all code, pretrained checkpoints, and evaluation framework under Apache 2.0 at this https URL Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.LG). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.31669 — 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 2026-10-07T22:14:27.531Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).