Skip to content

Add fasterrcnn documentation#295

Open
Chandraveersingh1717 wants to merge 11 commits intomlverse:mainfrom
Chandraveersingh1717:add-fasterrcnn-documentation
Open

Add fasterrcnn documentation#295
Chandraveersingh1717 wants to merge 11 commits intomlverse:mainfrom
Chandraveersingh1717:add-fasterrcnn-documentation

Conversation

@Chandraveersingh1717
Copy link
Contributor

Add Faster R-CNN object detection documentation example

Adds a new example article demonstrating object detection with model_fasterrcnn_resnet50_fpn(), following the fcnresnet article structure.

Features:

  • Documents latest output API (output$detections with boxes, labels, scores)
  • Shows how to load pretrained COCO weights and run inference
  • Demonstrates bounding box visualization with draw_bounding_boxes()
  • Explains eval() vs train() mode differences
  • Includes helper functions like coco_label() for class names

Tested and working with current torchvision API.

Demonstrate C++ function export with parameters for GSoC evaluation
Demonstrate C++ function export with parameters for GSoC evaluation
- Implement drawing lines between keypoints using connectivity matrix
- Add graphics::segments() to draw skeleton connections
- Update documentation with connectivity parameter and examples
- Add test for connectivity feature
- Add keypoint-skeleton.R example for human pose visualization
Add connectivity parameter to draw lines between keypoints with human pose example
Prevents 2D tensor from collapsing to 1D when indexing with single element.
Fixes convnext_detection model test failure.
C++ version clamps negative areas to 0, R version doesn't.
Correctness is tested separately, this only tests performance.
Documents latest output API, follows fcnresnet structure
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