outputs are all black images

zbf1991 / RRM

Reliability Does Matter: An End-to-End Weakly Supervised Semantic Segmentation Approach

94 stars 11 forks source link

outputs are all black images #6

Closed doulemint closed 4 years ago

doulemint commented 4 years ago

I am gonna ask a really dumb question. I downloaded the pre-train weight and also train it with pascal voc dataset. But the output from infer_RRM.py didn't give me color decoded and segmented images but a complete black images. I guess it might caused by the weight I loaded into the network, or some stupid mistakes. Does anyone have the same problem as I did....?

doulemint commented 4 years ago

I probably got it.... I simply changed cv2.imwrite to plt.imsave so it can be visualized....I guess there's other way to achieve the same result.