Skip to content

Add global_attention_mask to gen_kwargs in Seq2SeqTrainer.prediction_step #16485

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
sgugger merged 1 commit into from
Apr 5, 2022
Merged

Add global_attention_mask to gen_kwargs in Seq2SeqTrainer.prediction_step #16485

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
4 changes: 3 additions & 1 deletion src/transformers/trainer_seq2seq.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -163,9 +163,11 @@ def prediction_step(

if "attention_mask" in inputs:
gen_kwargs["attention_mask"] = inputs.get("attention_mask", None)
if "global_attention_mask" in inputs:
gen_kwargs["global_attention_mask"] = inputs.get("global_attention_mask", None)

# prepare generation inputs
# some encoder-decoder models can have varying encder's and thus
# some encoder-decoder models can have varying encoder's and thus
# varying model input names
if hasattr(self.model, "encoder") and self.model.encoder.main_input_name != self.model.main_input_name:
generation_inputs = inputs[self.model.encoder.main_input_name]
Expand Down