You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Standardize the set of sampling distributions. As of now we use the following:
'guassian' or 'normal': np.random.standard_normal
'uniform': np.random.uniform(-1, 1)
'None': np.random.rand
And also from csvd
'spixel': random sampling of rows
'sparse': np.sparse.rand
After we agree on the set of sampling distributions, we can incorporate them into
the sketch function to reduce code repetition and limit the set of distributions
on an algorithm by algorithm basis if needed (i.e. if rsvd and rdmd
should not have the same set of distributions available).
The text was updated successfully, but these errors were encountered:
Yes, that is a great idea to unify the sketch functionality across all functions. It will us also allow to easily add some additional sketching schemes in a future version.
Standardize the set of sampling distributions. As of now we use the following:
np.random.standard_normal
np.random.uniform(-1, 1)
np.random.rand
And also from
csvd
np.sparse.rand
After we agree on the set of sampling distributions, we can incorporate them into
the
sketch
function to reduce code repetition and limit the set of distributionson an algorithm by algorithm basis if needed (i.e. if
rsvd
andrdmd
should not have the same set of distributions available).
The text was updated successfully, but these errors were encountered: