Closed txfs1926 closed 4 years ago
BTW, when will the .pth.tar file to COCO-format panoptic predictions conversion script be released? Thanks.
Okay, I've written a script to achieve that.
Okay, I've written a script to achieve that.
Hey @txfs1926 I have been trying to visualize the Panoptic results with class labels but no luck so far, if possible can you please share your script.
Okay, I've written a script to achieve that.
Hi @txfs1926, could you share your transformation script with us? Would be awesome!
Okay, I've written a script to achieve that.
Hi @txfs1926, could you share your transformation script with us? Would be awesome!
Actually it's been nearly a year and I forgot a lot about this, but you might (might not) find some hints from my fork https://github.com/txfs19260817/seamseg or even from an application https://github.com/txfs19260817/seg-platform
Okay, I've written a script to achieve that.
Hi @txfs1926, could you share your transformation script with us? Would be awesome!
Actually it's been nearly a year and I forgot a lot about this, but you might (might not) find some hints from my fork https://github.com/txfs19260817/seamseg or even from an application https://github.com/txfs19260817/seg-platform
Thanks! Will have a look ;)
Hi, thanks for this great repo! The script
test_panoptic.py
only generates the semantic prediction map for each input image instead of panoptic one. How can I obtain the prediction maps in panoptic format, like those ground truth*_gtFine_instanceIds.png
of Cityscapes? Thanks!