Open zhou-rui1 opened 2 years ago
Hello, when I trying the Context_encoder code, it seems wrong with the pixel value of the mask region, looks like:
whereas, the original implementation code construct more clear region like this one:
do you know why?
Hello, when I trying the Context_encoder code, it seems wrong with the pixel value of the mask region, looks like:![image](https://user-images.githubusercontent.com/72313540/150689201-6c567fb6-dbc7-4491-b06b-e2652e12aa1b.png)
whereas, the original implementation code construct more clear region like this one:![image](https://user-images.githubusercontent.com/72313540/150689225-10ddb85b-453b-4c1d-8f20-e14ab4f06053.png)
do you know why?