Skip to content

[5680954,5620660@2][ONNX][Autocast] Update value info in converted graph #611

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

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[5680954,5620660@2][ONNX][Autocast] Update value info in converted graph #611

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
7 changes: 7 additions & 0 deletions modelopt/onnx/autocast/precisionconverter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -955,6 +955,13 @@ def _cleanup(self):
# Remove redundant casts
self._remove_redundant_casts()

# Update value_info in model's graph
self._update_value_info_in_graph()

def _update_value_info_in_graph(self):
for vi in self.model.graph.value_info:
vi.type.tensor_type.elem_type = self.value_info_map[vi.name].type.tensor_type.elem_type
Comment on lines +961 to +963
Copy link
Contributor

@galagam galagam Nov 30, 2025

Choose a reason for hiding this comment

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

Did you find the location where we update self.value_info_map instead of self.model.graph.value_info ?
Maybe graph outputs/inputs?

This fix is a little patchy IMO.


def _cleanup_no_consumer_nodes(self):
network_outputs = {o.name for o in self.model.graph.output}
nodes_to_remove = [
Expand Down
Loading