Tsai

When it fits

Use Tsai when you want a Python library for developing deep-learning workflows on sequential data and want access to its listed model families, task support and time-series evaluation utilities.

Who it is for

• Python developers working with time-series or sequential datasets.
• Practitioners building deep-learning experiments for classification, forecasting or regression.
• Researchers comparing recurrent, convolutional and transformer model approaches to time-series tasks.

Documented uses

• Develop deep-learning workflows for time-series classification.
• Build forecasting or regression experiments on sequential data.
• Explore listed CNN, RNN and transformer architectures for time-series problems.
• Use pipeline transforms and walk-forward cross-validation in time-series workflows.