fix _check_available in SBDInstanceSegmentationDataset

[Hakuyume]

======================================================================
ERROR: test_sbd_instance_segmentation_dataset (test_sbd_instance_segmentation_dataset.transplant_class.<locals>.C)  parameter: {'split': 'train'}
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/chainercv/tests/datasets_tests/sbd_tests/test_sbd_instance_segmentation_dataset.py", line 19, in setUp
    self.dataset = SBDInstanceSegmentationDataset(split=self.split)
  File "/usr/local/lib/python3.5/dist-packages/chainercv-0.8.0-py3.5-linux-x86_64.egg/chainercv/datasets/sbd/sbd_instance_segmentation_dataset.py", line 45, in __init__
    _check_available()
  File "/usr/local/lib/python3.5/dist-packages/chainercv-0.8.0-py3.5-linux-x86_64.egg/chainercv/datasets/sbd/sbd_instance_segmentation_dataset.py", line 22, in _check_available
    'so the dataset cannot be loaded.'
TypeError: category must be a Warning subclass, not 'str'

[knorth55]

Thank you so much !

[knorth55]

the comma at the end of this line causes this error.
I'm not sure, but RuntimeWarning does not matter according to https://github.com/chainer/chainercv/blob/master/chainercv/extensions/vis_report/detection_vis_report.py#L17-L22

[Hakuyume]

I see. I updated.