Skip to content
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

Jump to bottom

fix sequence parallel get labels #1352

Merged
merged 3 commits into from
Jul 10, 2024
Merged

fix sequence parallel get labels #1352

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2dd32af
fix get labels
tastelikefeet Jul 10, 2024
48a6c20
Merge commit 'a590aa2e77306f0aef56fe64da5de77d6db81a78' into fix/sequ…
tastelikefeet Jul 10, 2024
b194499
remove useless dependencies
tastelikefeet Jul 10, 2024
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
5 changes: 3 additions & 2 deletions swift/trainers/trainers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -202,12 +202,13 @@ def compute_loss(self, model, inputs, return_outputs=None):
else:
loss = outputs['loss'] if isinstance(outputs, dict) else outputs[0]

if labels is None:
labels = inputs['labels']

if self.sequence_parallel_size > 1:
from swift.trainers.xtuner import reduce_xtuner_sequence_parallel_loss
loss = reduce_xtuner_sequence_parallel_loss(loss, labels)

if labels is None:
labels = inputs['labels']
if self.is_encoder_decoder:
preds = outputs.logits.argmax(dim=2)[..., :]
labels = labels[..., :]
Expand Down
Loading