This benchmark evaluates the performance of various time series foundation models (TSFM) on a range of datasets and tasks. It is built using the benchopt framework, which provides a standardized way to compare different solvers and algorithms on a common set of problems.
The goal is to provide a benchmark that evaluate the models on:
- Classification
- Forecasting
- Anomaly Detection
- Event Detection
With diverse modalities (univariate, multivariate, EEG, etc.) and varying sequence lengths.
This benchmark can be run using the following commands:
$ pip install -U benchopt $ git clone https://github.com/benchopt/benchmark_tsfm $ benchopt run benchmark_tsfm
Apart from the problem, options can be passed to benchopt run, to restrict the benchmarks to some solvers or datasets, e.g.:
$ benchopt run benchmark_tsfm -s solver1 -d dataset2 --max-runs 10 --n-repetitions 10
Use benchopt run -h for more details about these options, or visit https://benchopt.github.io/api.html.