moabb
moabb copied to clipboard
NeuroTechX
Mother of All BCI Benchmarks
It could be interesting to add a metric (or function to compute) the score on the n% lowest performing subjects. In paradigms where we can get nearly perfect scores such...
There are no type hints in the evaluation classes, plus they use kwargs. Let's merge #720 #722 #723 then add type hints
> I would love to make one osfdataset with this... > Maybe instead of `LocalBIDSDataset`, you can create an `OsfDataset` class which inherits from `BaseBIDSDataset` and directly implements the `_download_subject`...
```suggestion # Data needs to be preloaded for the filtering step of paradigms raw.load_data() ``` Okay for now, but we need to remove this step. Full lazy in the future,...
With the new class `BaseBIDSDataset` (see #724), it becomes easier to include BIDS datasets in moabb. To further facilitate their integration, we should - [ ] integrate a first dataset...
https://mne.tools/stable/generated/mne.decoding.SlidingEstimator.html I feel that we don't need pseudo online splitter with this type of estimator
I am getting the error below when uisng the benchmark. It seems the error is in: compute_dataset_statistics() ``` Zhou2016-WithinSession: 100%|██████████| 4/4 [00:46
