Adds docs and Fixes Imports for ops and io

[oke-aditya]

Closes #2649 .

Adds docs for

• Image reading functions here

• Torchvision ops from boxes.py.

Some functions were missing in the __init__.py files, I have added those imports too.
Please review and do let me know if changes are needed.

cc @vfdev-5 @fmassa @pmeier

[codecov]

Codecov Report

Merging #2654 into master will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #2654   +/-   ##
=======================================
  Coverage   72.57%   72.57%           
=======================================
  Files          95       95           
  Lines        8249     8249           
  Branches     1309     1309           
=======================================
  Hits         5987     5987           
  Misses       1855     1855           
  Partials      407      407           

Impacted Files	Coverage Δ
torchvision/io/__init__.py	100.00% <ø> (ø)
torchvision/ops/__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 7f5b2c4...5ebdc3f. Read the comment docs.

[vfdev-5]

Hi @oke-aditya , sorry for delay, we discussed with @fmassa about this PR and we think that exposing decode_*/read_* from io module is not suitable as in future the idea is to provide a bit more generic functions like read and decode which could dispatch to those decode_*/read_* functions. Thus, user does not need to check the format and dispatch in his/her code. I'll open an issue to expose in details this idea.

For torchvision ops, they were not included in 0.3.0 release, but now we may consider them as more mature and make them public. Let me comment what we would to remove and we can merge this PR.

[vfdev-5]

@oke-aditya I left few comment. If you could handle that it would be perfect and we can merge this PR after that. Thanks !

[oke-aditya]

Ok @vfdev-5 sure, I will undo the changes from io module. I saw other PR being opened to revamp torchvision video reader and other APIs #2660 .

I will do changes as suggested 👍 and ask for a review once again

[vfdev-5]

LGTM ! Thanks @oke-aditya !