Madessa, Amanuel Hirpa

I had the same problem but fixed when I change train_image_classifier.py line 40 "clone_on_cpu" to True.

> > > I don't know what you actually did to generate your custom mask. > However, the generated mask image in evaluation step are the mask postprocessed by batch_postprocess_images...

After spending two days to solve this problem I finally found the cause and solved it. Let me explain it here with my poor English. The problem comes with converting...

i got same problem. How to fix this bug? thanks