[CI] Fix gluoncv tutorial under mxnet-mkl

[tqchen]

This is an error message generated by the mxnet-mkl image in the current master. The error was due to the operator amp_multicast is not implmented in the mxnet frontend(likely due to a mkl specialization).

http://ci.tvm.ai:8080/job/temp-ci-docker-staging/job/ci-stage2/6/execution/node/309/log/?consoleFull

Please see if we are interested in quick a fix, since we have docker upgrades depending on it. In the meanwhile, I will revert the docker image to mxnet for now until the next upgrade.

[tqchen]

cc @icemelon9 @shoubhik @anijain2305

[anijain2305]

@icemelon9 Do you know about this operator?

[icemelon]

It's weird that I cannot reproduce the error in my local environment with mxnet-mkl 1.6.0 and gluoncv.

[tqchen]

@icemelon9 could it due to that mxnet optionally enabled amp on the platform that support these ops?

[icemelon]

@tqchen Yes, that might be possible. I'll check it on a GPU instance.

[anijain2305]

Ping @icemelon9 If you get time.

[eric-haibin-lin]

I cannot reproduce the error either on Ubuntu GPU instance

from mxnet.contrib.quantization import quantize_model_mkldnn
  ImportError: cannot import name 'quantize_model_mkldnn'

for both mxnet==1.6.0 and mxnet-mkl==1.6.0