search

ryouchinsa / Rectlabel-support

RectLabel is an offline image annotation tool for object detection and segmentation.
https://rectlabel.com
515 stars 73 forks source link

Object detection model for grocery shelves #206

Closed joaourbano closed 2 years ago

joaourbano commented 2 years ago

Hi,

First of all, thank you for your great software.

I am currently annotating a few thousand images of grocery shelves. I have trained an object detection model using Create ML (YoloV2). I use this model with Rectalbel Pro to speed up my annotation process.

I understand that yoloV5 offers quite a few improvements and I would like to train a new model with it. Could you please help me out? Do you have a pre-trained model that I could download and use with RectLabel Pro?

Thank you

ryouchinsa commented 2 years ago

Thanks for writing the issue.

For the yolov5 models, you can download here, such as yolov5s.pt. https://github.com/ultralytics/yolov5/releases/tag/v6.2

To convert the .pt model to .mlmodel model, run the command. python export.py --weights yolov5s.pt --include coreml https://github.com/ultralytics/yolov5/issues/251

Then, you can load the yolov5s.mlmodel on RectLabel, process your images for automatic labeling. https://rectlabel.com/help#process_images_coreml

Please let us know your opinion.

ryouchinsa commented 2 years ago

Currently If you need our support to fix this problem, please let us know.