Fix to blau score

[SkafteNicki]

Before submitting

• Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

What does this PR do?

Implements the fix from
nltk/nltk@e562f62#diff-644ebc352f93bb9a63ff8edbc7292e137cf1adaf68d5dc4f1aef13f4de96bf20
so we again can test against latest nltk.

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #165 (67356ae) into master (935994c) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@            Coverage Diff            @@
##           master     #165     +/-   ##
=========================================
  Coverage   95.99%   95.99%             
=========================================
  Files         252       84    -168     
  Lines        7713     2572   -5141     
=========================================
- Hits         7404     2469   -4935     
+ Misses        309      103    -206     

Flag	Coverage Δ
Linux	78.73% <0.00%> (-0.04%)	⬇️
Windows	78.73% <0.00%> (-0.04%)	⬇️
cpu	95.99% <100.00%> (+<0.01%)	⬆️
gpu	?
macOS	95.99% <100.00%> (+<0.01%)	⬆️
pytest	95.99% <100.00%> (+<0.01%)	⬆️
python3.6	95.98% <100.00%> (+<0.01%)	⬆️
python3.8	95.99% <100.00%> (+<0.01%)	⬆️
python3.9	95.99% <100.00%> (+<0.01%)	⬆️
torch1.3.1	94.85% <100.00%> (+<0.01%)	⬆️
torch1.4.0	94.98% <100.00%> (+<0.01%)	⬆️
torch1.8.1	95.99% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
torchmetrics/functional/nlp.py	100.00% <100.00%> (ø)
__w/1/s/torchmetrics/classification/stat_scores.py
__w/1/s/torchmetrics/classification/__init__.py
...1/s/torchmetrics/retrieval/mean_reciprocal_rank.py
__w/1/s/torchmetrics/retrieval/retrieval_ndcg.py
__w/1/s/torchmetrics/classification/auroc.py
...hmetrics/functional/retrieval/average_precision.py
.../s/torchmetrics/classification/hamming_distance.py
.../1/s/torchmetrics/functional/retrieval/__init__.py
__w/3/s/torchmetrics/regression/psnr.py
... and 159 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 935994c...67356ae. Read the comment docs.