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
Hello !
I work mainly on computer vision and natural language processing.
Most of the time, we have a list of classes for example :
cat
dog
....
Each of this class is mapped with a unique ID.
cat => 0
dog => 1
...
This definition is unique accross the dataset and should be shared to every datarows.
I have two possibilities :
We store IDs in the dataset.
In this case we have everything we need to supply to the model the class id. But when we need to compute metrics we may struggle to find what is the real feature we are working on.
We store ClassNames in the dataset
In this case, we need to convert on the fly classname to Ids in order to feed our model. We need a global definition of ids in order to make sure we use the same everywhere.
=> We have to iterate over the whole dataset in order to check existing classes.
I'd like a way to persist this id/class_name definition somewhere in the dataset.
Is there a way to do this correctly or is it a feature request ?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hello !
I work mainly on computer vision and natural language processing.
Most of the time, we have a list of classes for example :
Each of this class is mapped with a unique ID.
This definition is unique accross the dataset and should be shared to every datarows.
I have two possibilities :
I'd like a way to persist this id/class_name definition somewhere in the dataset.
Is there a way to do this correctly or is it a feature request ?
Thanks,
Beta Was this translation helpful? Give feedback.
All reactions