I am using this implementation to reduce a dataset from 250,000 dimensions to 10 dimensions and it is taking almost 90 minutes while scikit's implementation takes only 4 to 5 minutes. However, scikit causes memory errors for reduction to more than 25 dimensions and that is why I would like to use this. Is there any way to make it faster because I need to also reduce to many more dimensions like 50, 100, 1000, 10000, 50000 etc. which will surely take much more time.
I am using this implementation to reduce a dataset from 250,000 dimensions to 10 dimensions and it is taking almost 90 minutes while scikit's implementation takes only 4 to 5 minutes. However, scikit causes memory errors for reduction to more than 25 dimensions and that is why I would like to use this. Is there any way to make it faster because I need to also reduce to many more dimensions like 50, 100, 1000, 10000, 50000 etc. which will surely take much more time.