[API][BUGFIX] remove extra dropout in BERT/RoBERTa

[eric-haibin-lin]

Description

dropout is already applied after the embedding is added with positional embedding

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 #1022 into master will increase coverage by 0.22%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1022      +/-   ##
==========================================
+ Coverage   88.05%   88.27%   +0.22%     
==========================================
  Files          67       67              
  Lines        6254     6252       -2     
==========================================
+ Hits         5507     5519      +12     
+ Misses        747      733      -14

Impacted Files	Coverage Δ
src/gluonnlp/model/bert.py	92.18% <100%> (-0.07%)	⬇️
src/gluonnlp/data/word_embedding_evaluation.py	96.94% <0%> (+0.76%)	⬆️
src/gluonnlp/data/batchify/batchify.py	95.27% <0%> (+0.78%)	⬆️
src/gluonnlp/data/transforms.py	82.58% <0%> (+1.12%)	⬆️
src/gluonnlp/data/dataset.py	99.2% <0%> (+1.58%)	⬆️
src/gluonnlp/data/corpora/wikitext.py	100% <0%> (+5.17%)	⬆️
src/gluonnlp/model/parameter.py	100% <0%> (+8%)	⬆️

[mli]

Job PR-1022/1 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1022/1/index.html

[mli]

Job PR-1022/2 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1022/2/index.html