Bodo Kaiser

From what should the network know if you want to detect an apple or a car? > Am 15.10.2017 um 21:58 schrieb 4F2E4A2E : > > Can you please explain...

You can’t say for sure what happens in the hidden layers but generally the network will only output what it was trained with. > Am 15.10.2017 um 22:03 schrieb 4F2E4A2E...

You should checkout „fine-tuning“. > Am 15.10.2017 um 23:05 schrieb 4F2E4A2E : > > Exactly, that's my honest opinion too, but my question is, how to train with the segmented...

Yes that is one application. Actually „fine-tuning“ is quite a common term in deep learning: http://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html It is also required by some of the semantic segmentation networks which use a...

@saeedizadi maybe has some?

I think you are right that binary segmentation only needs 1 output class / channel with 0 corresponding to first and 1 corresponding to second class. So something like: ```python...

Hey, yes, I think you are right. Could you test what happens if you add `F.log_softmax(labels)` [here](https://github.com/bodokaiser/piwise/blob/master/main.py#L101) and make a PR out of it?

You still need to download the data from the VOC project website or did you mean something else? > Am 15.05.2018 um 09:36 schrieb fantasyzl : > > Under the...

SegmentationClass or SegmentationClassObject contains the labels. The difference is that one encodes the classes (and maps them to a PNG color table) and the other one encodes the color codes...

@mshmoon there was another issue about this also see my disclaimer in the readme. For me I found that changing the u-net architecture to single conv's and using leaky relus...