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
When there are the same words in different modalities, we can't definetly restore phi from dataframe, and people have to do hacky things like saving several phi's for each modality.
We need fucntion like get_phi() in interface, but which returns array-like object, in which rows are ordered in accordance to word order in dictionary file (so I will be able to restore it definetly).
P.S. I know about get_phi().values, but it, in general case, randomly shuffles rows.
The text was updated successfully, but these errors were encountered:
ofrei
added a commit
to ofrei/bigartm
that referenced
this issue
Feb 13, 2017
@fedimser You are right, thanks for reporting this. May I ask you to review or test #759 and let me know if it it looks right? The change adds api ARTM.get_phi_dense that should solve this issue.
ofrei
added a commit
to ofrei/bigartm
that referenced
this issue
Feb 27, 2017
When there are the same words in different modalities, we can't definetly restore phi from dataframe, and people have to do hacky things like saving several phi's for each modality.
We need fucntion like get_phi() in interface, but which returns array-like object, in which rows are ordered in accordance to word order in dictionary file (so I will be able to restore it definetly).
P.S. I know about get_phi().values, but it, in general case, randomly shuffles rows.
The text was updated successfully, but these errors were encountered: