Error in Parsing PoolingND

[User-Ali]

I don't know much about Caffe. But, when I google to resolve this issue, I learned that there is no poolingnd in Caffe but some people were referring cudNN for it.

Now, can you please suggest how can I resolve this issue? PS : I don't have GPU so, no cudNN either.

Here is the layer:
layer { name: "pool1" type: "PoolingND" bottom: "conv1H" top: "pool1" poolingnd_param { pool: MAX kernel_h: 2 kernel_w: 2 kernel_d: 2 stride: 2 } }

[qxcv]

I don't think that error has anything to do with CuDNN. It looks like your model has an unusual layer from a third-party Caffe fork (maybe Caffe3D?). This converter only supports the default Caffe layers at the moment; patches are welcome if you'd like to extend it to support whatever fork you're using.

[User-Ali]

I wish I could but I don't have relevant knowledge. I could try to work on that if you provide some guidelines. Here is a related post on PoolingND.

It's used as a 3D Pooling layer but they have nothing mentioned about which Caffe version they used for it.