Custom Shufflenetv2 for extremely small and dense objects

[vuongnp]

Add new stride: 4 Total strides of the model: 4, 8, 16, 32, 64

[codecov[bot]]

Codecov Report

Merging #528 (4e25651) into main (226d8cc) will decrease coverage by 0.38%. The diff coverage is 56.31%.

@@            Coverage Diff             @@
##             main     #528      +/-   ##
==========================================
- Coverage   74.80%   74.42%   -0.38%     
==========================================
  Files          71       72       +1     
  Lines        4695     4798     +103     
  Branches      731      753      +22     
==========================================
+ Hits         3512     3571      +59     
- Misses        986     1022      +36     
- Partials      197      205       +8     

Flag	Coverage Δ
unittests	74.42% <56.31%> (-0.38%)	:arrow_down:

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

Files Changed	Coverage Δ
nanodet/model/backbone/shufflenetv2_dense.py	56.00% <56.00%> (ø)
nanodet/model/backbone/__init__.py	93.75% <66.66%> (+0.64%)	:arrow_up:

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more

[RangiLyu]

Thanks for your contribution! You can introduce the features of this model in the readme and the comments of the config to let other users know it better.