[BUGFIX] fix the in_channels of LN in qa script

[zheyuye]

Description

Fix the typo in question answering script of xlnet. The hidden units of xlnet large is 1024 which confilts with the fixed in_channels 768. see #17654.

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

• fix the typo in qa

Comments

@leezu @zburning
cc @dmlc/gluon-nlp-team

[codecov]

Codecov Report

Merging #1177 into master will increase coverage by 7.26%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1177      +/-   ##
==========================================
+ Coverage   80.08%   87.34%   +7.26%     
==========================================
  Files          80       80              
  Lines        7365     7365              
==========================================
+ Hits         5898     6433     +535     
+ Misses       1467      932     -535

Impacted Files	Coverage Δ
src/gluonnlp/data/word_embedding_evaluation.py	89.31% <0%> (+0.38%)	⬆️
src/gluonnlp/model/utils.py	80% <0%> (+0.76%)	⬆️
src/gluonnlp/model/train/embedding.py	87.17% <0%> (+2.56%)	⬆️
src/gluonnlp/data/datasetloader.py	85.34% <0%> (+2.58%)	⬆️
src/gluonnlp/embedding/token_embedding.py	91.79% <0%> (+2.82%)	⬆️
src/gluonnlp/vocab/elmo.py	96.66% <0%> (+3.33%)	⬆️
src/gluonnlp/data/registry.py	78.12% <0%> (+6.25%)	⬆️
src/gluonnlp/data/stream.py	88.42% <0%> (+7.36%)	⬆️
src/gluonnlp/model/transformer.py	91.66% <0%> (+10.89%)	⬆️
src/gluonnlp/model/bert.py	94.62% <0%> (+14.32%)	⬆️
... and 10 more

[mli]

Job PR-1177/1 is complete.
Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-1177/1/index.html

[leezu]

@sxjscience let's edit the commit message next time before merging. That will make it easier to prepare release notes.

For example:

[BUGFIX] Fix in_channels of LN (#1177)
Reference apache/mxnet#17654 apache/mxnet#17683

[sxjscience]

@leezu Okay.