[Feature Request] Add copy operator to MXNet to CoreML convertor

[ChaiBapchya]

As revealed in the issue #13965 currently only 12 operators are supported to convert from MXNet to CoreML.

    'FullyConnected' : _layers.convert_dense,
    'Activation'     : _layers.convert_activation,
    'SoftmaxOutput'  : _layers.convert_softmax,
    'Convolution'    : _layers.convert_convolution,
    'Pooling'        : _layers.convert_pooling,
    'Flatten'        : _layers.convert_flatten,
    'transpose'      : _layers.convert_transpose,
    'Concat'         : _layers.convert_concat,
    'BatchNorm'      : _layers.convert_batchnorm,
    'elemwise_add'   : _layers.convert_elementwise_add,
    'Reshape'        : _layers.convert_reshape,
    'Deconvolution'  : _layers.convert_deconvolution,

Based on the user request, we could add copy and similar commonly needed ops as well in this file

incubator-mxnet/tools/coreml/converter/_layers.py

@mxnet-label-bot add [Feature request]

[mxnet-label-bot]

Hey, this is the MXNet Label Bot. Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it. Here are my recommended labels: Feature

[frankfliu]

@mxnet-label-bot add [feature request]

[markproy]

Not sure how I add "+1" to this request, so I'm just adding this comment. I'd like to convert a ResNet-50 model created by MXNet under the hood of SageMaker's Image Classification built-in algo. Unfortunately, it uses "_copy" (and perhaps other unsupported layers). Thus, the converter is unable to create a coreML model.

[tapanh]

I have the exact same use case as @markproy above. Would be great to have this feature!

[mn033]

@markproy I have exactly the same issue with SageMaker Object Detection. Did you solve this problem in any way? (other converter etc.) ?

[tycoles]

Same problem here!