Skip to content

[Whisper] Fix audio feature device placement in encoder forward #22296

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

Closed
shenxiul wants to merge 1 commit into from
+5 −1
Closed

[Whisper] Fix audio feature device placement in encoder forward #22296

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: 5 additions & 1 deletion python/sglang/srt/models/whisper.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -458,8 +458,12 @@ def forward(
features.device, non_blocking=True
)

# Move features to the same device as model weights
device = self.encoder.conv1.weight.device
req_encoder_output = self.encoder(
features.to(dtype), encoder_position_ids, forward_batch
features.to(device=device, dtype=dtype),
encoder_position_ids.to(device),
forward_batch,
)
req_encoder_output = req_encoder_output.squeeze(0)
encoder_list.append(req_encoder_output)
Expand Down
Loading