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
The checkpoint storage in pyrevolve is a big numpy array, and the "pointers" that are given to checkpoint.save(ptr) and checkpoint.load(ptr) are in fact slices of that numpy array. A user who needs to store data that is not numpy will be out of luck. Not sure if this is really a problem in need of a fix, unless someone wants to do mixed-precision computing or checkpoint non-floating-point data. Open for discussion.
The text was updated successfully, but these errors were encountered:
The checkpoint storage in pyrevolve is a big numpy array, and the "pointers" that are given to
checkpoint.save(ptr)
andcheckpoint.load(ptr)
are in fact slices of that numpy array. A user who needs to store data that is not numpy will be out of luck. Not sure if this is really a problem in need of a fix, unless someone wants to do mixed-precision computing or checkpoint non-floating-point data. Open for discussion.The text was updated successfully, but these errors were encountered: