Skip to content

Populate last_hidden_state in the output of feature extraction model class #863

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
zhenchaoni merged 1 commit into from
Jun 10, 2026
+6 −0
Merged

Populate last_hidden_state in the output of feature extraction model class #863

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: 6 additions & 0 deletions src/winml/modelkit/models/winml/feature_extraction.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,4 +59,10 @@ def forward(self, **kwargs: Any) -> ModelOutput:
pooling handles 1-D and 2-D after raw[0].
"""
outputs = self._run_inference(self._format_inputs(**kwargs))
# WinMLEncoderDecoderModel expects its encoder sub-component to expose
# hidden states as "last_hidden_state". Alias the primary output when an
# encoder ONNX graph named it otherwise (e.g. "encoder_hidden_states").
# Appended last to preserve output[0] and the real ONNX output names.
if outputs and "last_hidden_state" not in outputs:
outputs["last_hidden_state"] = next(iter(outputs.values()))
return FeatureExtractionModelOutput(outputs)
Loading