Skip to content

Use a generator for instance segmentation masks #1604

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

Closed
wants to merge 3 commits into from
Closed

Use a generator for instance segmentation masks #1604

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
4bd6f74
Use a generator for instance segmentation masks
shntu Sep 30, 2025
0a0aa4e
Merge branch 'main' into sb/generate-seg-masks
shntu Sep 30, 2025
5a003ff
Merge branch 'main' into sb/generate-seg-masks
shntu Sep 30, 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
36 changes: 16 additions & 20 deletions inference/core/models/instance_segmentation_base.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -245,28 +245,24 @@ def make_response(
"""
responses = []
for ind, (batch_predictions, batch_masks) in enumerate(zip(predictions, masks)):
predictions = []
for pred, mask in zip(batch_predictions, batch_masks):
if class_filter and not self.class_names[int(pred[6])] in class_filter:
# TODO: logger.debug
Copy link
Contributor Author

Choose a reason for hiding this comment

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

Would a debug log be necessary? This PR deletes this TODO since it didn't seem especially helpful, but I can also add logs if needed to the generator.

continue
# Passing args as a dictionary here since one of the args is 'class' (a protected term in Python)
predictions.append(
InstanceSegmentationPrediction(
**{
"x": pred[0] + (pred[2] - pred[0]) / 2,
"y": pred[1] + (pred[3] - pred[1]) / 2,
"width": pred[2] - pred[0],
"height": pred[3] - pred[1],
"points": [Point(x=point[0], y=point[1]) for point in mask],
"confidence": pred[4],
"class": self.class_names[int(pred[6])],
"class_id": int(pred[6]),
}
)
predictions_gen = (
InstanceSegmentationPrediction(
**{
"x": pred[0] + (pred[2] - pred[0]) / 2,
"y": pred[1] + (pred[3] - pred[1]) / 2,
"width": pred[2] - pred[0],
"height": pred[3] - pred[1],
"points": [Point(x=point[0], y=point[1]) for point in mask],
"confidence": pred[4],
"class": self.class_names[int(pred[6])],
"class_id": int(pred[6]),
}
)
for pred, mask in zip(batch_predictions, batch_masks)
if not class_filter or self.class_names[int(pred[6])] in class_filter
)
response = InstanceSegmentationInferenceResponse(
predictions=predictions,
predictions=predictions_gen,
image=InferenceResponseImage(
width=img_dims[ind][1], height=img_dims[ind][0]
),
Expand Down
4 changes: 2 additions & 2 deletions inference/models/yolact/yolact_instance_segmentation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -272,11 +272,11 @@ def make_response(
"""
responses = [
InstanceSegmentationInferenceResponse(
predictions=[
predictions=(
Copy link
Contributor

Choose a reason for hiding this comment

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

Isn't it converted to list before rendering? Not sure if it help with memory usage.

Copy link
Contributor Author

@shntu shntu Oct 2, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

Currently, there are a handful of enterprise users who import InferencePipeline and run it without actually using any of the HTTP responses.

I'm closing this PR because it indeed does not solve the problem - the memory usage is much higher on instance segmentation execution, and reducing the memory used by these masks is a very small fraction of it (would require a much larger change to actually reduce the memory pressure on the Jetson).

InstanceSegmentationPrediction(**p)
for p in batch_pred
if not class_filter or p["class_name"] in class_filter
],
),
image=InferenceResponseImage(
width=img_dims[i][1], height=img_dims[i][0]
),
Expand Down