Good day,
Our team labeled 1000+ images of the new game ball (blue and red). We ended up not needing it, but if your team would like to use it, you can download the collection (1 GB) below.
https://drive.google.com/file/d/1F-qmSxL0R-z7l7R3ozSfZic0yWLF9QCw/view?usp=sharing
Class names: [redball, blueball]
Images: 1000+
Label-type: rectangle
Hope this helps!
Jack
Programming Lead - Texas Torque (FRC 1477)
Any way to upload this back to supervisely to combine datasets?
Yes,
If you go to Import, you can choose the plugin “Supervisely”. This should let you then upload the entire folder of images (untarred).
Hope this helps!
So I was wondering if you guys had this dataset in a YoloV5 format? I tried uploading to Supervisely and running the “Supervisely to YoloV5 format” converter, but it gives me a folder with only 209 images vs the 1000+ in the original dataset.
Interesting. I went to the project on my side and converted it. Kindly try to download it from here to see if this fixes your issue?
Best,
Jack
You’re awesome. Thanks!!
I’m going to try and get this dataset trained into a yolov5 model and share my results.
Oh, and one piece of information to be aware of: as it is the culmination of five videos, there are five different .tar files inside the file I sent. You will likely have to join these together in your code (should be able to merge easily in Python) as I couldn’t find a way to posthumously join them in Supervise.ly.
Glad to have helped,
Jack
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.