why the test result of the compressed gaugan trained by me is not as expected? but the training result of supernet is right.
yuanlunxi opened this issue · comments
_
Sorry, I do not quite understand your problem? Could you please elaborate more. Which dataset are you using?
I used the dataset of myself,the dataset is correct, because the result of training process is correct.
in the fast gan-compression process, when i export the compressed the model from the supernet,should i use the super_mobile_sapde or mobile_spade?
the test output of head0 of compressed gaugan is very huge,
so is there wrong step cause by me in the export process?
I bet there is a problem with input data normalization in your code (preprocessing of your dataset for compression)
I bet there is a problem with input data normalization in your code (preprocessing of your dataset for compression)
the input is normalized
I find the output of
is very huge
i find the problem, when the self.training in the SuperSynchronizedBatchNorm2d is set True ,the result of test is right. but can anyone explain this ?
- what was the size of each image in your dataset? 256*256?
- are all the images the same size?
- what values did you change to solve this issue