[Feature] Add Machine translation estimator in api

[liuzh47]

Description

Implementation of machine translation GNMT and transformer estimator.

Checklist

Essentials

• PR's title starts with a category (e.g. [BUGFIX], [MODEL], [TUTORIAL], [FEATURE], [DOC], etc)
• Changes are complete (i.e. I finished coding on this PR)
• All changes have test coverage
• Code is well-documented

Changes

• Feature1, tests, (and when applicable, API doc)
• Feature2, tests, (and when applicable, API doc)

Comments

• If this change is a backward incompatible change, why must this change be made.
• Interesting edge cases to note here

cc @dmlc/gluon-nlp-team

[codecov]

Codecov Report

Merging #1156 into master will increase coverage by 3.18%.
The diff coverage is 26.22%.

@@            Coverage Diff             @@
##           master    #1156      +/-   ##
==========================================
+ Coverage   70.58%   73.77%   +3.18%     
==========================================
  Files          72       76       +4     
  Lines        6970     7317     +347     
==========================================
+ Hits         4920     5398     +478     
+ Misses       2050     1919     -131

Impacted Files	Coverage Δ
src/gluonnlp/estimator/__init__.py	100% <100%> (ø)
...nlp/estimator/machine_translation_event_handler.py	22.22% <22.22%> (ø)
...p/estimator/machine_translation_batch_processor.py	28.57% <28.57%> (ø)
...luonnlp/estimator/machine_translation_estimator.py	54.54% <54.54%> (ø)
src/gluonnlp/data/translation.py	26.35% <0%> (-73.65%)	⬇️
src/gluonnlp/model/train/cache.py	25.58% <0%> (-72.1%)	⬇️
src/gluonnlp/model/transformer.py	31.73% <0%> (-54.81%)	⬇️
src/gluonnlp/data/batchify/language_model.py	43.92% <0%> (-52.34%)	⬇️
src/gluonnlp/model/translation.py	20.31% <0%> (-51.57%)	⬇️
src/gluonnlp/embedding/evaluation.py	40.33% <0%> (-51.27%)	⬇️
... and 45 more

[mli]

Job PR-1156/1 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/1/index.html

[liuzh47]

@eric-haibin-lin

[eric-haibin-lin]

Why does grad acc handler API require batch_size?

[liuzh47]

It needs batch_size to rescale. See here https://github.com/dmlc/gluon-nlp/blob/master/scripts/machine_translation/train_transformer.py#L320

[mli]

Job PR-1156/2 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/2/index.html

[mli]

Job PR-1156/3 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/3/index.html

[mli]

Job PR-1156/4 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/4/index.html

[mli]

Job PR-1156/5 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/5/index.html

[mli]

Job PR-1156/6 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/6/index.html

[mli]

Job PR-1156/7 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/7/index.html

[mli]

Job PR-1156/8 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1156/8/index.html

[chenw23]

@eric-haibin-lin Do you think this pull request should be fixed on CI and then be merged?