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
Thank you for this great work.
I'm just a bit confused—are the results with or without face detection?
When I trained MobileNet with the original dataset, it reached 97.98%. However, when I used the dataset with face detection preprocessing, it only reached 97.45%.
The text was updated successfully, but these errors were encountered:
For training process, the dataset is assumed already handled by face detection. Thus the result is without face detection.
So the 97.98% is printed during traing, and 97.45% is your own testing process result? As the result is rather close, maybe just some difference in image processing like Reproduce the results #110, or flip process?
Thank you for this great work.
I'm just a bit confused—are the results with or without face detection?
When I trained MobileNet with the original dataset, it reached 97.98%. However, when I used the dataset with face detection preprocessing, it only reached 97.45%.
The text was updated successfully, but these errors were encountered: