-
Notifications
You must be signed in to change notification settings - Fork 2.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[MO] Move reverse input channels before mean/scale #9182
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me, thanks a lot!
@@ -59,11 +59,14 @@ Framework-agnostic parameters: | |||
--reverse_input_channels | |||
Switch the input channels order from RGB to BGR (or | |||
vice versa). Applied to original inputs of the model |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@nosovmik, why do we decide to stay "Applied to original inputs of the model..." instead of "Applied to user's input data..."?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is a legacy behavior, and it is consistent with model's documentation. E.g. if model is RGB and states that means (120, 130, 140) shall be applied, but user has BGR input - in this case user will need to --mean_values(120,130,140) --reverse_input_channels
instead of converting something in the mind and have --mean_values(140,130,120) --reverse_input_channels
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let us merge it and continue this discussion offline
No description provided.