Revisiting the Generic Transformer: Deconstructing a Strong Baseline for Time Series Foundation Models
This work provides a transparent, reproducible baseline for time series foundation model research, addressing the difficulty in attributing improvements to architectural innovations versus data engineering in the field.
The authors investigated whether a standard patch Transformer architecture could achieve state-of-the-art zero-shot forecasting performance for time series data using a straightforward training protocol, and found that it does while identifying key performance drivers through comprehensive ablation studies.
The recent surge in Time Series Foundation Models has rapidly advanced the field, yet the heterogeneous training setups across studies make it difficult to attribute improvements to architectural innovations versus data engineering. In this work, we investigate the potential of a standard patch Transformer, demonstrating that this generic architecture achieves state-of-the-art zero-shot forecasting performance using a straightforward training protocol. We conduct a comprehensive ablation study that covers model scaling, data composition, and training techniques to isolate the essential ingredients for high performance. Our findings identify the key drivers of performance, while confirming that the generic architecture itself demonstrates excellent scalability. By strictly controlling these variables, we provide comprehensive empirical results on model scaling across multiple dimensions. We release our open-source model and detailed findings to establish a transparent, reproducible baseline for future research.