Expose frame-rate and cache to video datasets

[fmassa]

Those arguments have been separated from #1303 to facilitate review.

cc @stephenyan1231

[stephenyan1231]

LGTM. Ready to go

[codecov-io]

Codecov Report

Merging #1356 into master will decrease coverage by 0.42%.
The diff coverage is 25%.

@@            Coverage Diff             @@
##           master    #1356      +/-   ##
==========================================
- Coverage   65.34%   64.92%   -0.43%     
==========================================
  Files          75       76       +1     
  Lines        5878     5953      +75     
  Branches      902      911       +9     
==========================================
+ Hits         3841     3865      +24     
- Misses       1765     1817      +52     
+ Partials      272      271       -1

Impacted Files	Coverage Δ
torchvision/datasets/ucf101.py	25% <0%> (ø)	⬆️
torchvision/datasets/hmdb51.py	27.65% <0%> (ø)	⬆️
torchvision/datasets/kinetics.py	34.78% <50%> (ø)	⬆️
torchvision/io/__init__.py	100% <0%> (ø)	⬆️
torchvision/io/_video_opt.py	27.02% <0%> (ø)
torchvision/transforms/transforms.py	81.56% <0%> (+0.58%)	⬆️

---

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 a6a926b...164eeb3. Read the comment docs.