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
Since these names are used later to name dictionaries/processes, it would be nice if we directly used the name of the attribution we used. I'm envisioning something like:
# The version in a config fileattributions= ["DDeepLift", "DIntegratedGradients"]
# The version in a scriptattributions= [DDeepLift, DIntegratedGradients]
We could both get attributions by name, or get a name from an attribution object. This could be extended to things like the processor used, etc. This could make naming conventions more consistent across runs.
The text was updated successfully, but these errors were encountered:
In the current tutorials and set ups for doing attribution and evaluation, we name things via proxy by using dictionaries like:
Since these names are used later to name dictionaries/processes, it would be nice if we directly used the name of the attribution we used. I'm envisioning something like:
We could both get attributions by name, or get a name from an attribution object. This could be extended to things like the processor used, etc. This could make naming conventions more consistent across runs.
The text was updated successfully, but these errors were encountered: