laboroai / laborotomato Goto Github PK

Thanks to the author for your contribution, LabororTomato has been helpful to my research. How can I cite it in my paper?

Thank you very much for your work, which is very helpful to me, but I can't download the dataset from your download link, could you please share Google Cloud disk or other download links。

I see that both in train.json and test.json have the following format for bounding boxes:
"annotations" : [ { ... "image_id" : 0, "bbox" : [ 1475.0769599999999, 1567.11744, 425.2046400000004, 394.8134400000001 ], ... }
I first thought these numbers would correspond to opposing corners of a rectangle but that doesn't seem to be correct. I would be grateful if you could help me find out how those bounding boxes work.

hi i tried mmdet v2 and mmdetv3 but i could not run the model correctly. can you please give me the steps and especially the config files. becasue as far as i understand, pre-trained model is trained in mmdetv2 but your example config files are mmdetv3 so there is a confusion. they are not compatible. please give clear explenation for how to run pre-trained model correctly, maybe code snippet etc.