cpu version :param input_tensor: :param clip_value: :param variance: :param device: :return:

gpu compatible version :param input_tensor: variable :param clip_value: clipping value of 2-norm :param variance: variance of gau

gpu compatible version :param input_tensor: variable :param clip_value: clipping value of 2-norm :param variance: variance of gau

return appr -125-125 BGR

create a group of models, namely feature extractor, aggregator and discriminator, for each feature group

Translate bounding boxes to fit within the cropped area of an image. This method is mainly used together with image cropping. This method tra

Flip bounding boxes accordingly. The bounding boxes are expected to be packed into a two dimensional tensor of shape :math:`(R, 4)`, where :m

Forward Faster R-CNN. Scaling paramter :obj:`scale` is used by RPN to determine the threshold to select small objects, which are goin

Forward Region Proposal Network. Here are notations. * :math:`N` is batch size. * :math:`C` channel size of the input.

Forward the chain. We assume that there are :math:`N` batches. Args: x (Variable): 4D image variable. rois (