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
I think using the default graph is wrong, because it enforces an assumption of 1 graph per script, which may be fine in the beginning, but as time goes on this will be a nagging limitation if all scripts are written to assume 1 graph. Note that this extends to the process as well, so you can't import any code which uses its own graph.
Therefore I am left to conclude that using the default graph is the true anti-pattern.
The text was updated successfully, but these errors were encountered:
Or, maybe I'm reading the documentation wrong?
http://www.tensorflowpatterns.org/patterns/use-default-graph.html
I think using the default graph is wrong, because it enforces an assumption of 1 graph per script, which may be fine in the beginning, but as time goes on this will be a nagging limitation if all scripts are written to assume 1 graph. Note that this extends to the process as well, so you can't import any code which uses its own graph.
Therefore I am left to conclude that using the default graph is the true anti-pattern.
The text was updated successfully, but these errors were encountered: