Skip to content

Feat/support wavlm large #1889

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

Open
antoinelaurent wants to merge 3 commits into
base: develop
Choose a base branch
from
Open

Feat/support wavlm large #1889

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a4edf58
Adding support for WAVLM_LARGE architecture
antoinelaurent Jul 4, 2025
a176f11
fix a typo
antoinelaurent Jul 4, 2025
687e491
Update SSeRiouSS.py
antoinelaurent Jul 4, 2025
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
20 changes: 17 additions & 3 deletions src/pyannote/audio/models/segmentation/SSeRiouSS.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -226,7 +226,12 @@ def num_frames(self, num_samples: int) -> int:
"""

num_frames = num_samples
for conv_layer in self.wav2vec.feature_extractor.conv_layers:

# Torchaudio's WAVLM_LARGE has an extra .model wrapper
wav2vec_model = getattr(self.wav2vec, "model", self.wav2vec)
fe = wav2vec_model.feature_extractor

for conv_layer in fe.conv_layers:
num_frames = conv1d_num_frames(
num_frames,
kernel_size=conv_layer.kernel_size,
Expand All @@ -251,8 +256,12 @@ def receptive_field_size(self, num_frames: int = 1) -> int:
Receptive field size.
"""

# Torchaudio's WAVLM_LARGE has an extra .model wrapper
wav2vec_model = getattr(self.wav2vec, "model", self.wav2vec)
fe = wav2vec_model.feature_extractor

receptive_field_size = num_frames
for conv_layer in reversed(self.wav2vec.feature_extractor.conv_layers):
for conv_layer in reversed(fe.conv_layers):
receptive_field_size = conv1d_receptive_field_size(
num_frames=receptive_field_size,
kernel_size=conv_layer.kernel_size,
Expand All @@ -275,8 +284,13 @@ def receptive_field_center(self, frame: int = 0) -> int:
receptive_field_center : int
Index of receptive field center.
"""

# Torchaudio's WAVLM_LARGE has an extra .model wrapper
wav2vec_model = getattr(self.wav2vec, "model", self.wav2vec)
fe = wav2vec_model.feature_extractor

receptive_field_center = frame
for conv_layer in reversed(self.wav2vec.feature_extractor.conv_layers):
for conv_layer in reversed(fe.conv_layers):
receptive_field_center = conv1d_receptive_field_center(
receptive_field_center,
kernel_size=conv_layer.kernel_size,
Expand Down