discretecoder

Unified panoptic output error

1

I am getting the following error. I am training on 8+1 thing and 32 total classes and my weights are also matching the number of classes but I am still...

Training on own dataset

1

I am trying to train the network on a custom dataset where I have the rgb images and panoptic labels in png format. I am following your tip and trying...

Testing on own data or on the test images in cityscape

1

I am trying to run the test code to infer from the test images but I get an error that the annotation file does not exist for the data. i...

I trained the network on 38 classes. but when i test the model, i tried printing the np.unique(pan) from the output and it has class ids greater than 38 so...

In the coco experiment, pred.json file only stores the predictions from one image. should it not store all the predictions from the entire val set?? thanks

Hi, Thanks for the great work, what is the panos output format I assume the panos in the output is supposed to have indices in the range of class ids...

hi what does this parameter do

when i try to save unified panoptic output, the width of the output image is not the same as the input image. eg: if i have an imput image of...

Hi, I wondering if you have higher level features from cityscapes training like the imagenet pretrained weights to use on a similar dataset? thanks

Greetings, I have a question, if I have panoptic labels where I use 5 different shades od blue for cars for example like in cityscapes panoptic labels, how should i...