You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank you for your great work on this project. I am doing another research work where your parts dataset of Cityscape and Pascal will be of great help.
I have obtained the train and val set from this repo but I am a little lost on the process to obtain the test set. Would you be willing to share the test set with me so I can generate my evaluation stats?
I would really appreciate it if you could share the test set images and ground truth for parts. My problem statement is different so I will need to do some ETL on the test set so will be a huge help if you can share the set with me. If not, just images and references to the evaluation server will be great.
Thanks a lot! Much appreciate your help.
The text was updated successfully, but these errors were encountered:
Regarding Cityscapes Panoptic Parts, the test set of the original Cityscapes dataset is not publicly available (by the original authors), thus we could not annotate it.
Regarding PASCAL Panoptic Parts, we create the part-aware panoptic annotations based on annotations from PASCAL-Context and PASCAL-Parts. These datasets provide annotations only for the 10103 training and validation images, so the test images are not annotated.
In any case, I think you could use the validation sets for your "test" experiments as the majority of works are doing in the field.
Thank you for your great work on this project. I am doing another research work where your parts dataset of Cityscape and Pascal will be of great help.
I have obtained the train and val set from this repo but I am a little lost on the process to obtain the test set. Would you be willing to share the test set with me so I can generate my evaluation stats?
I would really appreciate it if you could share the test set images and ground truth for parts. My problem statement is different so I will need to do some ETL on the test set so will be a huge help if you can share the set with me. If not, just images and references to the evaluation server will be great.
Thanks a lot! Much appreciate your help.
The text was updated successfully, but these errors were encountered: