How to obtain the segmentation map when the input text contains only one text label instead of a whole set of labels of a dataset?

[jinx2018]

Hi! Thanks for releasing the codes!

Based on the codes, in the inference stage to obtain the segmentation mask, the input texts contains a whole set of labels of some dataset. And I wonder if the text contains only one text label, how to get the segmentation mask?

Looking forward to your reply!