Alexandre Gramfort comments

Results 758 comments of


                                            Alexandre Gramfort

Allow EEGclassifier in cropped to work with gridsearch cv or pipeline

option b) seems the way to go for me > Message ID: ***@***.***> >

Implement GCNs-Net Model for EEG Motor Imagery Signal Decoding in EEGdecoder Repository

yes it would be great to grow the model zoo here ! @Buddies-as-you-know can you give us a hand ?

Rename Dataset Classes

you have now EEGClassifier and EEGRegressor. I would try to be consistent so EEGDataset so an EEGClassifier will be fed by an EEGDataset.

Substituting Custom Layers with Native PyTorch Equivalents

+1 to migrate to more standard implementations but each change deserves a check of non-regression on a relevant task. Maybe MOABB can help with this? > Message ID: github.com> >

[ENH] Create functions to read Epochs and Evoked Neuroscan formats

it was introduced when we created the EvokedArray class. > Message ID: ***@***.***> >

[ENH] Create functions to read Epochs and Evoked Neuroscan formats

make an EvokedArray as done eg by read_evoked_besa > Message ID: ***@***.***> >

[ENH] Create functions to read Epochs and Evoked Neuroscan formats

Evoked only knows nave so not the number of dropped. The Epochs do know this. Yes you could use the comment for this... > Message ID: ***@***.***> >

Make ECoG and sEEG size defaults more realistic

@alexrockhill ok for you?

Make ECoG and sEEG size defaults more realistic

both would work for me but maybe a better default is enough? > Message ID: ***@***.***> >

Replace percent format with format specifiers (f-strings)

see >>> basename = "toto" >>> name_ext = ".txt" >>> "%s_div%%i%s" % (basename, name_ext) 'toto_div%i.txt' >>> f"{basename}_div%i{name_ext}" 'toto_div%i.txt' so f"{basename}_div%i{name_ext}" should be the correct replacement. Message ID: ***@***.***> >