convert error pytorch0.2

[yja1]

torch.version
'0.2.0_4'
ValueError: Unknown layer type: Negate, known types: ['Slice', 'MaxPool2d', 'Add', 'PReLU', 'ELU', 'MulConstant', 'Dropout', 'Addmm', 'BatchNorm', 'Concat', 'UpsamplingBilinear2d', 'Tanh', 'Softmax', 'ConvNd', 'Cmax', 'data', 'View', 'AddConstant', 'Sigmoid', 'LeakyReLU', 'AvgPool2d', 'Threshold']

when convert pytorch Faceboxes-->caffe

[starimeL]

The official version of caffe don't support several layers in Faceboxes. You can try ncnn:)

[KaleidoZhouYN]

The official version of caffe don't support several layers in Faceboxes. You can try ncnn:)

is that because Faceboxes use the CReLU activate?

[yja1]

I try to convert https://github.com/lxg2015/faceboxes/blob/master/networks.py

[starimeL]

@yja1 the network is similar to the Faceboxes here

[starimeL]

@KaleidoZhouYN Also some of other layers like Permute, etc.

[KaleidoZhouYN]

@starimeL

@KaleidoZhouYN Also some of other layers like Permute, etc.

so can we just ignore the conversion of permute layer and artificially add caffe layer in prototxt to reach the same result about dim transpose?