I found I can circumvent some of the initial gpuArray errors by setting ks.ops.GPU= 0; which allows me to at least view my data (preprocessed beforehand via highpass filtering and common average referencing prior to writing to a .bin file). Some pre-processing steps inside of kilosort (e.g. whitening filter) seem dependent on GPU operations.
Must i look towards Kilosort4 in python, else get a GPU?
I found I can circumvent some of the initial
gpuArray
errors by settingks.ops.GPU= 0;
which allows me to at least view my data (preprocessed beforehand via highpass filtering and common average referencing prior to writing to a .bin file). Some pre-processing steps inside of kilosort (e.g. whitening filter) seem dependent on GPU operations.Must i look towards Kilosort4 in python, else get a GPU?