Fix find_unused_parameters in PyTorch 1.8 for BasicVSR

[ckkelvinchan]

As discussed in #273, training of BasicVSR halted due to the settings of find_unused_parameters. This PR fixes the error in PyTorch1.8 by removing the modification of find_unused_parameters during training. find_unused_parameters is kept True throughout training.

[codecov]

Codecov Report

Merging #290 (f34b49d) into master (3cfdd8f) will decrease coverage by 0.00%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master     #290      +/-   ##
==========================================
- Coverage   80.54%   80.54%   -0.01%     
==========================================
  Files         162      162              
  Lines        8091     8090       -1     
  Branches     1192     1192              
==========================================
- Hits         6517     6516       -1     
  Misses       1429     1429              
  Partials      145      145              

Flag	Coverage Δ
unittests	80.54% <66.66%> (-0.01%)	⬇️

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

Impacted Files	Coverage Δ
mmedit/models/restorers/basicvsr.py	68.88% <66.66%> (-0.35%)	⬇️

---

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 3cfdd8f...f34b49d. Read the comment docs.

[innerlee]

@nbei please confirm that the changes is okay.