submission errors (Xulin Chen (大胖胖墩), Ye Zhang (pluto_charon) ¶
By: crunch on April 5, 2022, 12:40 p.m.
Hi, we noticed that the last submission by visilab failed (model file was missing) - please contact us for more details or if you need more help (applies to all participants with or without failed submissions :-) ).
Update2: vunos submission failed due to (still the same error): ModuleNotFoundError: No module named 'vuno.detection..modeling'
Update3 @sleepygray: File til-score.json was not produced (needs to be created even if not used for L1 evaluation)
Update4@sleepygray: The container was killed as it exceeded the memory limit of 16g
Update5@vuno: "RuntimeError: Found no NVIDIA driver on your system" This is strange, usually this would indicate something wrong with the (aws) node, but also maybe something is wrong with the installed cuda version in your docker?
Update6@大胖胖墩 The file produced at detected-lymphocytes.json is not valid - please see the example on submission details webpage.
Update 8th April: @gdph: ValueError: images do not match segmentation_mask.paste(mask,(x,y,x+seg_size_x,y+seg_size_y)) File "/venv/lib/python3.8/site-packages/PIL/Image.py", line 1584, in paste self.im.paste(im, box)
@大胖胖墩. The lymphocytes are still invalid, I assume because the coords are still numpy and saved with <> in the json. x: 8.39248150281592 x float: 8.39248150281592
Update 22nd April: please be aware that the docker will not have internet access, so trying to download models, etc. will either result in an error or a hanging job terminated after the time limit ends.
Best, Witali