[Enhancement] only use the best model for inference for bert classification

[eric-haibin-lin]

Description

@szha

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

Merging #716 into master will increase coverage by 0.1%.
The diff coverage is n/a.

@@            Coverage Diff            @@
##           master     #716     +/-   ##
=========================================
+ Coverage   90.53%   90.63%   +0.1%     
=========================================
  Files          65       65             
  Lines        5991     5991             
=========================================
+ Hits         5424     5430      +6     
+ Misses        567      561      -6

Impacted Files	Coverage Δ
src/gluonnlp/data/dataloader.py	88.79% <0%> (+5.17%)	⬆️

[codecov]

Codecov Report

❗ No coverage uploaded for pull request head (best-acc@b8042c4). Click here to learn what that means.
The diff coverage is n/a.

[mli]

Job PR-716/1 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-716/1/index.html

[mli]

Job PR-716/2 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-716/2/index.html