Skip to content

Fix BNR 2.0 inference alignment error with input signal padding #15390

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
ko3n1g merged 1 commit into from
Feb 17, 2026
Merged

Fix BNR 2.0 inference alignment error with input signal padding #15390

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
17 changes: 13 additions & 4 deletions nemo/collections/audio/models/maxine/bnr.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -214,10 +214,19 @@ def forward(self, input_signal):
)
)

if input_signal.shape[-1] % SUPPORTED_INPUT_ALIGN_SAMPLES != 0:
raise ValueError("Input samples must be a multiple of {}".format(SUPPORTED_INPUT_ALIGN_SAMPLES))

return self.seasr.forward(x0=input_signal)
# Pad input to nearest multiple of SUPPORTED_INPUT_ALIGN_SAMPLES
original_length = input_signal.shape[-1]
remainder = original_length % SUPPORTED_INPUT_ALIGN_SAMPLES
if remainder != 0:
pad_length = SUPPORTED_INPUT_ALIGN_SAMPLES - remainder
input_signal = F.pad(input_signal, (0, pad_length))
output = self.seasr.forward(x0=input_signal)

# Trim output back to original length
if remainder != 0:
output = output[..., :original_length]

return output

def training_step(self, batch, batch_idx):
if isinstance(batch, dict):
Expand Down
Loading