Yubin Wang

@igorbarinov what's the difference between instance nomalization and batchnorm with batchsize = 1 ?

@igorbarinov got it. instance normalization only normalized in spatial not cross feature map and batch

@bucktoothsir I haven't tried. but batchsize = 1 still normalize cross the feature maps. instance normalization only do normalization within the feature map plane. As @yusuketomoto say, batchsize >=2 have...

@bucktoothsir sorry . After i checked the paper, i found batchsize = 1 is instance normalization.

@ttoinou batch normalization is across instances, however, when batch size = 1, there is only one instance. You can also implement multi batch training with normaliztion within instance.

@vermapratyush you may run second time. The first run cost longer time.

Yes, there are so many hyper-parameters began from the seminal work "neural algorithm for artistic stylization". Currently most re-implementation follows uniform lambda_style weights. I have done lots for experiments based...

@DavieHR @littlekobe I think that doesn't make a lot of diference. This VGG mean is just around 120.

@dbrazey the `num_nodes(trt_only)` = 0 says that your environment is something wrong. You may try tensorflow 1.15 with tensorrt included, or try nvidia tensorflow docker container.

@sudevbohra you may pay attention to the input image of conv1, such as whether need substract mean value, and the input range, also notice the metal weight is [outChannel][height][width][inputChannel], which...