Skip to content

Modify custom BatchNorm layer such that the mean and variance variables have the same rank whether train=False or train=True. #838

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.

Sign up for GitHub

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

Merged
copybara-service merged 1 commit into from
Jan 30, 2026
+3 −3
Merged

Modify custom BatchNorm layer such that the mean and variance variables have the same rank whether train=False or train=True. #838

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions init2winit/model_lib/deepspeech.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -401,17 +401,17 @@ def __call__(self, inputs, input_paddings=None, train=False):

if train:
mask = 1.0 - padding
sum_v = jnp.sum(inputs * mask, axis=reduce_over_dims, keepdims=True)
sum_v = jnp.sum(inputs * mask, axis=reduce_over_dims, keepdims=False)
count_v = jnp.sum(
jnp.ones_like(inputs) * mask, axis=reduce_over_dims, keepdims=True)
jnp.ones_like(inputs) * mask, axis=reduce_over_dims, keepdims=False)

count_v = jnp.maximum(count_v, 1.0)
mean = sum_v / count_v

sum_vv = jnp.sum(
(inputs - mean) * (inputs - mean) * mask,
axis=reduce_over_dims,
keepdims=True)
keepdims=False)

var = sum_vv / count_v

Expand Down