Output attention takes an s

[sgugger]

Fixes #6902

Stas would have come with a nice Perl magic command but I did a regex search (output_attention[^s]) to fix all those misspelled args. In the process, I noticed a few examples were missing a line so added that too.

[codecov]

Codecov Report

Merging #6903 into master will increase coverage by 0.29%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #6903      +/-   ##
==========================================
+ Coverage   79.30%   79.59%   +0.29%     
==========================================
  Files         157      157              
  Lines       28853    28853              
==========================================
+ Hits        22882    22966      +84     
+ Misses       5971     5887      -84     

Impacted Files	Coverage Δ
src/transformers/configuration_auto.py	93.18% <ø> (ø)
src/transformers/configuration_utils.py	96.64% <ø> (+0.67%)	⬆️
src/transformers/modelcard.py	85.18% <ø> (ø)
src/transformers/modeling_auto.py	78.73% <ø> (ø)
src/transformers/modeling_distilbert.py	97.84% <ø> (ø)
src/transformers/modeling_encoder_decoder.py	92.00% <ø> (ø)
src/transformers/modeling_tf_auto.py	66.86% <ø> (ø)
src/transformers/modeling_tf_distilbert.py	64.47% <ø> (-34.36%)	⬇️
src/transformers/modeling_tf_utils.py	87.29% <ø> (+0.32%)	⬆️
src/transformers/modeling_utils.py	87.50% <ø> (ø)
... and 23 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 485da72...e8fd79c. Read the comment docs.

[LysandreJik]

Great, thanks!