fix training with single class dataset

[fcakyon]

fixes #4971

[codecov]

Codecov Report

Merging #4973 (fb74386) into master (bca2766) will increase coverage by 1.14%.
The diff coverage is 45.00%.

@@            Coverage Diff             @@
##           master    #4973      +/-   ##
==========================================
+ Coverage   65.59%   66.73%   +1.14%     
==========================================
  Files         255      259       +4     
  Lines       20036    20270     +234     
  Branches     3407     3447      +40     
==========================================
+ Hits        13142    13527     +385     
+ Misses       6188     6017     -171     
- Partials      706      726      +20     

Flag	Coverage Δ
unittests	66.69% <45.00%> (+1.14%)	⬆️

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

Impacted Files	Coverage Δ
mmdet/apis/inference.py	38.18% <4.76%> (-5.12%)	⬇️
mmdet/datasets/lvis.py	14.86% <10.00%> (+0.86%)	⬆️
mmdet/datasets/api_wrappers/coco_api.py	83.33% <83.33%> (ø)
mmdet/datasets/api_wrappers/__init__.py	100.00% <100.00%> (ø)
mmdet/datasets/coco.py	46.25% <100.00%> (-0.06%)	⬇️
mmdet/datasets/utils.py	100.00% <100.00%> (ø)
mmdet/models/roi_heads/base_roi_head.py	85.71% <100.00%> (ø)
mmdet/models/detectors/cornernet.py	94.87% <0.00%> (-5.13%)	⬇️
mmdet/models/dense_heads/corner_head.py	73.72% <0.00%> (-1.88%)	⬇️
mmdet/models/detectors/__init__.py	100.00% <0.00%> (ø)
... and 12 more

---

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 a616886...fb74386. Read the comment docs.