European Distributed Deep Learning (EDDL) library. A general-purpose library initially developed to cover deep learning needs in healthcare use cases within the DeepHealth project.
This code that used to work up to v0.8a now (v0.8.1a) leads to a segmentation fault:
layer out = Activation(Dense(l, num_classes), "softmax");
The workaround is to replace it with:
layer out = Softmax(Dense(l, num_classes));
I don't know if the intention is that the old way should still work: if so, it should be fixed. If not, it should raise some exception, rather than resulting in a segmentation fault.
Note that the old idiom still shows up in some examples.
This code that used to work up to v0.8a now (v0.8.1a) leads to a segmentation fault:
The workaround is to replace it with:
I don't know if the intention is that the old way should still work: if so, it should be fixed. If not, it should raise some exception, rather than resulting in a segmentation fault.
Note that the old idiom still shows up in some examples.