Postdam eval_segmentation.py returning ValueError: Unknown Dataset

[COvert96]

Traceback (most recent call last):
File "eval_segmentation.py", line 117, in my_app
raise ValueError("Unknown Dataset {}".format(model.cfg.dataset_name))
ValueError: Unknown Dataset potsdam

After working through a number of environment issues after conda env create -f environment.yml such as pillow throwing DLL errors, I reinstalled most packages with pip.

Now I get this error, which seems to be because the model.cfg.dataset_name is only being checked against cocostuff and cityscapes? How are other people evaluating/reproducing the postdam results?

[mamadinho]

I have the same issue, still can't do evaluation on the Potsdam dataset.

[mamadinho]

Turns out you just need to comment the visualization code in the eval code