Support mmcv bricks in flops compuation

[innerlee]

Fix #675

Running the script in that issue

before:

Sequential(
  (0): Conv2d(30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  (0): Conv2d(30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  0.008 M, 50.000% Params, 0.531 GFLOPs, 100.000% FLOPs, 
  (0): Conv2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 100.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(0.0 M, 0.000% Params, 0.0 GFLOPs, 0.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  0.016 M, 100.000% Params, 1.062 GFLOPs, 100.000% FLOPs, 
  (0): Conv2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)

after:

Sequential(
  (0): Conv2d(30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  (0): Conv2d(30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  0.016 M, 100.000% Params, 1.062 GFLOPs, 100.000% FLOPs, 
  (0): Conv2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)
Sequential(
  0.016 M, 100.000% Params, 1.062 GFLOPs, 100.000% FLOPs, 
  (0): Conv2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
  (1): ConvTranspose2d(0.008 M, 50.000% Params, 0.531 GFLOPs, 50.000% FLOPs, 30, 30, kernel_size=(3, 3), stride=(2, 2), output_padding=(1, 1), bias=False)
)

[codecov]

Codecov Report

Merging #715 (5735233) into master (e30dc4f) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #715   +/-   ##
=======================================
  Coverage   64.90%   64.91%           
=======================================
  Files         135      135           
  Lines        7870     7872    +2     
  Branches     1396     1396           
=======================================
+ Hits         5108     5110    +2     
  Misses       2497     2497           
  Partials      265      265           

Flag	Coverage Δ
unittests	64.91% <100.00%> (+<0.01%)	⬆️

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

Impacted Files	Coverage Δ
mmcv/cnn/utils/flops_counter.py	93.18% <100.00%> (+0.06%)	⬆️

---

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 e30dc4f...3084b0e. Read the comment docs.