onnx / onnx/models

Possible bug in YOLOv3 model?

Open
#434 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug
Dominant language
Jupyter Notebook
Stars
9.8k
Forks
1.6k
PR merge metrics
No merged PRs in 30d

Description

In yolov3-10.onnx (from https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/yolov3 ), the post-processing part includes 6 loops. The trip counts are 13, 13, 26, 26, 52 ,52, respectively.

The loops uses scan outputs. The problem is, by ONNX definition, loop output is like "[1, 2, 3, 4...TripCount].
However, for this model, it expects the output to be like "[0, 1, 2, 3, .. TripCount-1]"

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with yolov3-10.onnx from the linked ONNX Models path and inspect the six post-processing loops with trip counts 13, 13, 26, 26, 52, and 52. Compare their scan outputs with ONNX Loop output semantics, especially whether the model expects zero-based values. Done means the loop-output indexing is consistent with the ONNX definition.

Written by the indexing model from the issue text.

Assessment

Domain
computer-vision, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.