feat: Add recognition metric (exact match)

[charlesmindee]

This PR implements ExactMatch metric (word-level accuracy) for recognition task
Any feedback is welcome !

[codecov]

Codecov Report

Merging #110 (9daed8d) into main (169fbac) will decrease coverage by 0.36%.
The diff coverage is 85.18%.

@@            Coverage Diff             @@
##             main     #110      +/-   ##
==========================================
- Coverage   97.79%   97.43%   -0.37%     
==========================================
  Files          28       29       +1     
  Lines         908      935      +27     
==========================================
+ Hits          888      911      +23     
- Misses         20       24       +4     

Flag	Coverage Δ
unittests	97.43% <85.18%> (-0.37%)	⬇️

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

Impacted Files	Coverage Δ
doctr/utils/metrics.py	84.61% <84.61%> (ø)
doctr/utils/__init__.py	100.00% <100.00%> (ø)

---

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 169fbac...c0c1614. Read the comment docs.

[fg-mindee]

Thanks for the PR! I added a few comments. Besides, especially if we don't ignore accents or case, we could actually do this with TF and compare the output before decoding?

[charlesmindee]

Thanks for the PR! I added a few comments. Besides, especially if we don't ignore accents or case, we could actually do this with TF and compare the output before decoding?

I think so, but it would require tensor manipulations to compare row-wise both tensors. What I was doing previously was translating with the dictionary to a string tensor and I concatenated the columns to have a vector of words, but I think we can compare the number sequences at a row-level without translating the outputs to save time.

[fg-mindee]

We might want later to convert this to a TF-only computation, but it will do for now! Thanks for the edits 🙏