Where can I find the class definitions? The Cityscapes dataset has 30 classes now.

[Chan0081]

I'm trying to use the pretrained model with id 48333157_3 for Semantic Segmentation. It was trained on Cityscapes dataset. I tried the demo.py and the prediction shape is (19, H, W), which means there are 19 classes predicted.
The paper says

Cityscapes is an urban egocentric street-view dataset
with high-resolution images (1024 x 2048 pixels). It contains
2975 images for training, 500 images for validation
and 1525 images for testing with a total of 19 classes.

However, the Cityscapes has now 30 classes.
So where can I find the definition of the 19 classes used when the model was trained?

[Chan0081]

Done. The class names can be found in predicor.metadata:
stuff_classes=['road', 'sidewalk', 'building', 'wall', 'fence', 'pole', 'traffic light', 'traffic sign', 'vegetation', 'terrain', 'sky', 'person', 'rider', 'car', 'truck', 'bus', 'train', 'motorcycle', 'bicycle']
BTW, it would be really nice if the model could be trained on the current Cityscapes dataset with 30 classes :)