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

Generate: get the correct beam index on eos token #18851

Merged
merged 2 commits into from
Sep 5, 2022
Merged

Generate: get the correct beam index on eos token #18851

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
297bbf7
Get the correct beam index on eos token
gante Sep 1, 2022
78b2dc1
Fix test as well
gante Sep 1, 2022
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
2 changes: 1 addition & 1 deletion src/transformers/generation_beam_search.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -259,7 +259,7 @@ def process(
continue
if beam_indices is not None:
beam_index = beam_indices[batch_beam_idx]
beam_index = beam_index + (next_index,)
beam_index = beam_index + (batch_beam_idx,)
Copy link
Member Author

@gante gante Sep 1, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

batch_beam_idx -> beam idx considering all batches (between num_beams*(batch_idx-1) and (num_beams*batch_idx)-1)

next_index -> beam idx for the current batch (between 0 and num_beams-1)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah yeah great point!

else:
beam_index = None

Expand Down
2 changes: 1 addition & 1 deletion tests/generation/test_generation_beam_search.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,7 +172,7 @@ def cut_expected_tensor(tensor):
input_ids[correct_idx].tolist(), beam_scorer._beam_hyps[batch_idx].beams[0][1].tolist()
)
self.parent.assertListEqual(
expected_beam_indices + [next_indices[batch_idx, 1].item()],
expected_beam_indices + [correct_idx],
torch.tensor(beam_scorer._beam_hyps[batch_idx].beams[0][2]).tolist(),
)

Expand Down