Transformer architectures for time series

How attention-based models adapt to sequential, non-stationary data — and where they underperform simpler baselines.

This lesson is part of our Model Design track. The full walkthrough, including code notebooks and worked examples, is being prepared by our research team and will publish directly to this page.

In the meantime, our research blog covers related methodology in depth, and our Telegram community is a good place to ask questions about this topic directly.