Any plans to update with Instance normalization? about chainer-fast-neuralstyle HOT 11 OPEN

On the same system, is the Torch implementation or this implementation faster? better? Anyone have comparisons?

from chainer-fast-neuralstyle.

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

from chainer-fast-neuralstyle.

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

from chainer-fast-neuralstyle.

@austingg Have you tried it? I think it does the same thing with batch normalization when you set batchsize = 1

from chainer-fast-neuralstyle.

@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 some problem, i do all my experiments with batchsize = 1

from chainer-fast-neuralstyle.

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

from chainer-fast-neuralstyle.

Are you sure ? According to the paper that would mean that using batchsize=1 yields better results than batchsize >= 2

from chainer-fast-neuralstyle.

@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.

from chainer-fast-neuralstyle.

From the paper we can see the effect is greatly improved. Anybody have the implement in python, I can test it.

from chainer-fast-neuralstyle.

@logic1988 : Yes they updated their project yesterday with instance normalization with torch ! https://github.com/DmitryUlyanov/texture_nets
@austingg : Great so there's still some hope to get the training faster by fixing the batchsize issue (I guess that loading images, resizing them and giving them to the GPU would be faster in batch, no ? there would be less CPU <-> GPU communication)

from chainer-fast-neuralstyle.

There are a great Torch Implementation released by jcjohnson a few days ago.
https://github.com/jcjohnson/fast-neural-style

In addition to the original method, the new method inspired by the the paper [Instance Normalization: The Missing Ingredient for Fast Stylization] has been implemented.

The latter seems to have a better result and faster speed than the former.

It said that:

These models use the same architecture as those used in our paper, except with half the number of filters per layer and with instance normalization instead of batch normalization. Using narrower layers makes the models smaller and faster without sacrificing model quality.

How to modify the corresponding code in this project? @yusuketomoto

from chainer-fast-neuralstyle.