Pierre Guetschel

see #452

Implementing a paradigm that allows doing a "cropped" epoching (in the sense of braindecode: https://braindecode.org/stable/auto_examples/plot_bcic_iv_2a_moabb_cropped.html) would be relatively simple with the new modular paradigms. We would just need to implement...

Confusion matrices

1

In the multi-classes evaluations, it would be nice to also get the confusion matrices in the results.

Improve parallelisation of evaluations

3

After discussions at the braindecode code sprint and following up on #460, I think we should break down the evaluations into something like that: ```python class BaseEvaluation: def __init__( self,...

Save results immediately

4

Closes #419 Instead of waiting for the evaluation of a whole dataset to finish before saving the results, we could save them as soon as they are computed. Beneficial in...

Allow `paradigm.get_data()` to return non-loaded Epochs

1

Since PR #408, we theoretically have the possibility to return non-loaded (i.e. `preload=False`) epochs and raws from the cache. However, because the paradigms' `get_data` concatenates epochs, they are loaded anyway...

update benchmarks CLI

2

Hi, what would you say about using [Fire](https://github.com/google/python-fire) to launch benchmarks from a CLI instead of the `run.py` script?

`Ofner2017(executed=True)` can't set montage and fails

1

Hi, new bug encountered: ```python from moabb.datasets import Ofner2017 Ofner2017(executed=True).get_data([1]) ``` ``` --------------------------------------------------------------------------- ValueError Traceback (most recent call last) Cell In[10], line 1 ----> 1 Ofner2017(executed=True).get_data([1]) File ~/miniconda3/envs/motor_embedding_benchmark/lib/python3.10/site-packages/moabb/datasets/base.py:113, in BaseDataset.get_data(self,...

Hi, the `init` argument of the `UMAP` class is not rendered as it should in [the doc](https://umap-learn.readthedocs.io/en/latest/api.html#umap.umap_.UMAP)