seq2seq/run_eval.py can take decoder_start_token_id

[sshleifer]

Also: batch_decode: document kwargs for autocomplete

[codecov]

Codecov Report

Merging #5949 into master will decrease coverage by 0.17%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master    #5949      +/-   ##
==========================================
- Coverage   78.69%   78.51%   -0.18%     
==========================================
  Files         146      146              
  Lines       26214    26214              
==========================================
- Hits        20628    20581      -47     
- Misses       5586     5633      +47     

Impacted Files	Coverage Δ
src/transformers/tokenization_utils_base.py	93.60% <50.00%> (ø)
src/transformers/modeling_tf_electra.py	26.02% <0.00%> (-69.52%)	⬇️
src/transformers/generation_utils.py	96.82% <0.00%> (-0.29%)	⬇️
src/transformers/modeling_tf_distilbert.py	98.79% <0.00%> (+33.89%)	⬆️

---

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 95d1962...450d753. Read the comment docs.

[sgugger]

Docstring of batch_decode could be better, but I'll soon do a swipe other this file so that's ok for now.