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
because it doesn't account for the broadcasting (i.e. it gives a value of 2 in the above example, when it should be 2 * 3 * 4).
broadcasting from (2,) to (2, 3, 4) makes sense here, but elsewhere, "broadcasting" may starts with the rightmost dimension, i.e. broadcast (4,) to (2, 3, 4)
The mean/variance calculations are incorrect, which means the inputs are not normalized correctly. E.g.
gives output
instead of the correct normalized output (
[0, 0, 0, 0]
).The basic issue is that this calculation is incorrect:
keras/keras/src/layers/normalization/batch_normalization.py
Lines 310 to 314 in efaaf85
because it doesn't account for the broadcasting (i.e. it gives a value of 2 in the above example, when it should be 2 * 3 * 4).
See #19818 for more discussion/background.
The text was updated successfully, but these errors were encountered: