Skip to content

⚡ Bolt: Optimize Perception NMS with torchvision#2

Open
harvatechs wants to merge 1 commit intomainfrom
bolt-perception-optimization-9809566504949198455
Open

⚡ Bolt: Optimize Perception NMS with torchvision#2
harvatechs wants to merge 1 commit intomainfrom
bolt-perception-optimization-9809566504949198455

Conversation

@harvatechs
Copy link
Owner

This PR optimizes the Non-Maximum Suppression (NMS) algorithm in the perception module by leveraging the highly optimized torchvision.ops.nms implementation when available.

Changes:

  • Modified agents/perception.py to conditionally import nms from torchvision.ops.
  • Used torchvision.ops.nms if available, falling back to the existing pure Python implementation otherwise.
  • Retained the box_iou helper function for the fallback implementation.

Performance:

  • Benchmark showed a ~28x speedup (432ms vs 15ms) for processing 2000 boxes.
  • This significantly reduces the latency of the perception pipeline, especially for object detection.

Verification:

  • Verified correctness by comparing outputs of both implementations.
  • Verified fallback mechanism works correctly.

PR created automatically by Jules for task 9809566504949198455 started by @harvatechs

- 💡 What: Replaced custom NMS implementation with `torchvision.ops.nms` in `agents/perception.py`.
- 🎯 Why: The custom Python implementation was a performance bottleneck ($O(N^2)$ in Python loops).
- 📊 Impact: Measured ~28x speedup (432ms -> 15ms for 2000 boxes on CPU).
- 🔬 Measurement: Verified with benchmark script comparing execution time and correctness.
- 🛡️ Compatibility: Added fallback to custom implementation if `torchvision` is not available.

Co-authored-by: harvatechs <191946902+harvatechs@users.noreply.github.com>
@google-labs-jules
Copy link

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.


For security, I will only act on instructions from the user who triggered this task.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant