onnx / onnx/models

At Readme of YOLOv3, Value types of original image size need to be float

Open
#671 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

Bug Report

Which model does this pertain to?

YoloV3

Describe the bug

Each of hosted models(YOLOv3, YOLOv3-12, YOLOv3-12-int8) need the original image size as the second input.
These Onnx Models expect float values for the second input, but the preprocessing steps sets it as dtype=np.int32, this causes asserting by type error.
It need to be np.float32 or 'float32' for the parameter dtype on this last line, same as Tiny Yolov3.

I will make a pull request for this issue soon.

Reproduction instructions

System Information

OS Platform and Distribution:
Windows 11 22H2

ONNX version:
not needed for reproducing this issue

Backend/Runtime version:
onnxruntime 1.18.0
numpy 1.26.0

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

Open validated/vision/object_detection_segmentation/yolov3/README.md and read the preprocessing steps, especially the final dtype setting for the original image size. Compare it with the linked Tiny YOLOv3 README. Done means the YOLOv3 README uses the float dtype expected by the hosted models instead of int32.

Written by the indexing model from the issue text.

Assessment

Tech stack
numpy, python
Domain
documentation
Issue type
Documentation
Difficulty
1/5
Estimated time
Under an hour
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.