Bug when test solov2

1

Hello,
I am trying to test GitHub - WXinlong/SOLO: SOLO and SOLOv2 for instance segmentation, ECCV 2020 & NeurIPS 2020.
I have do training but :
I have the following bug when I launch
python tools/test.py configs/solov2/solov2_r50_fpn_8gpu_3x.py demo/epoch_36.pth --json_out output --eval segm
:
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>] 30/30, 5.1 task/s, elapsed: 6s, ETA: 0sTraceback (most recent call last):
File “tools/test.py”, line 282, in
main()
File “tools/test.py”, line 273, in main
results2json(dataset, outputs, args.json_out)
File “c:\solo-master\mmdet\core\evaluation\coco_utils.py”, line 224, in results2json
json_results = det2json(dataset, results)
File “c:\solo-master\mmdet\core\evaluation\coco_utils.py”, line 153, in det2json
for i in range(bboxes.shape[0]):#changed by SA
AttributeError: ‘NoneType’ object has no attribute ‘shape’

and when I run the demo I have always 0 instances
Please help me !
Best regards

2

I have instances but always only one. That I want is to segment the leaflets of a leaf but if the leaf is simple it works but if the leaf is compound all the leaflets are in the same instance, an idea ?

3

I don’t quite understand the last sentence, but the previous error is raised if no bboxes are detected and they are thus None. My guess is that the code you are using isn’t properly guarding against this case and thus fails.