[Model] Add pre-trained BERT model on OpenWebText

[eric-haibin-lin]

Description

Add BERT base model pre-trained on OpenWebText.

Source	GluonNLP	google-research/bert	google-research/bert
Model	bert_12_768_12	bert_12_768_12	bert_24_1024_16
Dataset	openwebtext_book_corpus_wiki_en_uncased	book_corpus_wiki_en_uncased	book_corpus_wiki_en_uncased
SST-2	95.3	93.5	94.9
RTE	73.6	66.4	70.1
QQP	72.3	71.2	72.1
SQuAD 1.1	91.0/84.4	88.5/80.8	90.9/84.1
STS-B	87.5	85.8	86.5
MNLI-m/mm	85.3/84.9	84.6/83.4	86.7/85.9

Special thanks to @davisliang for preparing the dataset and running the pre-training :)

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

[codecov]

Codecov Report

❗ No coverage uploaded for pull request head (webtext@6769db4). Click here to learn what that means.
The diff coverage is n/a.

[codecov]

Codecov Report

Merging #806 into master will increase coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #806      +/-   ##
==========================================
+ Coverage   90.38%   90.39%   +<.01%     
==========================================
  Files          66       65       -1     
  Lines        6378     6280      -98     
==========================================
- Hits         5765     5677      -88     
+ Misses        613      603      -10

Impacted Files	Coverage Δ
src/gluonnlp/model/bert.py	99.27% <ø> (ø)	⬆️
src/gluonnlp/data/utils.py	74.14% <ø> (ø)	⬆️
src/gluonnlp/data/stream.py	85.56% <0%> (-4.06%)	⬇️
src/gluonnlp/data/dataloader.py
src/gluonnlp/utils/files.py	45.09% <0%> (+8.73%)	⬆️

[mli]

Job PR-806/1 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/1/index.html

[mli]

Job PR-806/2 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/2/index.html

[mli]

Job PR-806/3 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/3/index.html

[mli]

Job PR-806/4 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/4/index.html

[mli]

Job PR-806/5 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/5/index.html

[mli]

Job PR-806/6 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/6/index.html

[mli]

Job PR-806/7 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/7/index.html

[mli]

Job PR-806/8 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-806/8/index.html