Skip to content

Fix emb loss scale #6597

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
merged 3 commits into from
Nov 13, 2025
+1 −2
Merged

Fix emb loss scale #6597

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
64a461e
fix docs
tastelikefeet Nov 10, 2025
cffed2f
fix loss
tastelikefeet Nov 13, 2025
a25a5d4
Merge commit 'f031e4ea0bd7457b3304033229c3a87d53c68b9e' into fix/emb-…
tastelikefeet Nov 13, 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
3 changes: 1 addition & 2 deletions swift/plugin/loss.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -446,7 +446,7 @@ def infonce_loss(outputs, labels, loss_scale=None, num_items_in_batch=None, **kw
similarity_matrix = torch.cat(logits_list, dim=1)
# temperature scaling and CE
similarity_matrix = similarity_matrix / temperature
loss = nn.CrossEntropyLoss()(similarity_matrix, labels) / world_size # avoid duplicate
loss = nn.CrossEntropyLoss()(similarity_matrix, labels)
else:
all_tensors = []
for tensor in split_tensors:
Expand Down Expand Up @@ -499,7 +499,6 @@ def infonce_loss(outputs, labels, loss_scale=None, num_items_in_batch=None, **kw
# next positive is neg+1
length += tensor.size(0) - 1
loss /= len(split_tensors)
loss /= world_size # avoid duplicate
return loss


Expand Down
Loading