I have a question. Is the training data used by the author the same as the one released? I trained my own model, but it has overexposure issues, for example: areas that were originally bright in the image become very bright. However, the official model doesn't have this problem. I'd like to ask how to solve it?
I have a question. Is the training data used by the author the same as the one released? I trained my own model, but it has overexposure issues, for example: areas that were originally bright in the image become very bright. However, the official model doesn't have this problem. I'd like to ask how to solve it?