[DataCollatorForLanguageModeling] fix labels

[patil-suraj]

This PR fixes #6211. Only set ignore index (-100) when pad_token_id is not None

@sgugger

[sgugger]

Thanks for fixing this!

[codecov]

Codecov Report

Merging #6213 into master will decrease coverage by 0.38%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master    #6213      +/-   ##
==========================================
- Coverage   79.66%   79.27%   -0.39%     
==========================================
  Files         146      146              
  Lines       26582    26583       +1     
==========================================
- Hits        21176    21075     -101     
- Misses       5406     5508     +102     

Impacted Files	Coverage Δ
src/transformers/data/data_collator.py	96.58% <50.00%> (-0.84%)	⬇️
src/transformers/modeling_tf_distilbert.py	64.40% <0.00%> (-34.39%)	⬇️
src/transformers/tokenization_roberta.py	76.71% <0.00%> (-21.92%)	⬇️
src/transformers/tokenization_utils_base.py	86.43% <0.00%> (-7.42%)	⬇️
src/transformers/tokenization_transfo_xl.py	38.73% <0.00%> (-3.76%)	⬇️
src/transformers/tokenization_utils_fast.py	92.14% <0.00%> (-2.15%)	⬇️
src/transformers/tokenization_openai.py	82.57% <0.00%> (-1.52%)	⬇️
src/transformers/tokenization_bert.py	90.86% <0.00%> (-0.46%)	⬇️
src/transformers/tokenization_utils.py	90.00% <0.00%> (-0.41%)	⬇️
src/transformers/file_utils.py	80.05% <0.00%> (-0.26%)	⬇️
... and 3 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 9996f69...a0d4291. Read the comment docs.

[LysandreJik]

LGTM! Thanks @patil-suraj