remove _fork

[zhreshold]

Description

Fix shared Dataloader after removing _fork() in upstream mxnet. apache/mxnet#12554

Checklist

Essentials

• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage
• Code is well-documented
• Interesting edge cases to note here

[szhengac]

The mxnet version in env needs to be updated.

[szha]

Let's hold onto this change so that we can release 0.4.1 w/ mxnet 1.3.0.

[mli]

Job PR-353/2 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-353/2/index.html

[mli]

Job PR-353/3 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-353/3/index.html

[mli]

Job PR-353/7 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-353/7/index.html

[codecov]

Codecov Report

Merging #353 into master will decrease coverage by 1%.
The diff coverage is 31.25%.

@@            Coverage Diff             @@
##           master     #353      +/-   ##
==========================================
- Coverage   62.47%   61.46%   -1.01%     
==========================================
  Files          84       84              
  Lines        7088     7069      -19     
  Branches     1122     1116       -6     
==========================================
- Hits         4428     4345      -83     
- Misses       2357     2436      +79     
+ Partials      303      288      -15

Flag	Coverage Δ
#PR353	61.46% <31.25%> (-0.85%)	⬇️
#PR370	?
#master	?
#py2	61.26% <31.25%> (-1%)	⬇️
#py3	61.23% <31.25%> (-0.07%)	⬇️

[szha]

Suggested change

	from mxnet.test_utils import download
	from mxnet.gluon.utils import download

[szha]

Suggested change

	import mxnet as mx
	import os
	import mxnet as mx

[szha]

Suggested change

	download(url_format.format(filename))
	download(url_format.format(filename), path=os.path.join('tests', 'data', filename))

[szha]

Suggested change

	rec_dataset = gluon.data.RecordFileDataset(filename)
	rec_dataset = gluon.data.RecordFileDataset(os.path.join('tests', 'data', filename))

[mli]

Job PR-353/18 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-353/18/index.html