Error rendering potsdam label

[qAp]

Hi
When RENDER_DATA is set to True in potsdam.py, an AssertionError arising from the data having a wrong shape is returned:

~/git_repos/IIC/IIC/datasets/segmentation/potsdam.py in _prepare_test(self, index, img, label)
    336 
    337     if RENDER_DATA:
--> 338       render(label, mode="label", name=("test_data_label_pre_%d" % index))
    339 
    340     # convert to coarse if required, reindex to [0, gt_k -1], and get mask

~/git_repos/IIC/IIC/utils/segmentation/render.py in render(data, mode, name, colour_map, offset, out_dir)
     23       return
     24     else:
---> 25       assert (len(data.shape) == 3)
     26   else:
     27     assert (mode == "mask" or mode == "matrix" or mode == "preds")

AssertionError: 

Loading a potsdam example's ground truth mat file this way:

assert label_path.exists()
label = sio.loadmat(label_path)['gt']
assert label.dtype==np.int32
label.shape

returns an array, label, with shape (200, 200), which has length 2.

[xu-ji]

I recommend writing your own code to render the data. I think that flag is quite old and possibly unsupported.

See my reply here.

[qAp]

k thanks