:param inp: Multi person ground truth heatmap input (17 ch) Each channel contains multiple joints. :return: out: Gaussian augmented output. V

:param inp: Single person ground truth heatmap input (17 ch) Each channel contains one joint. :return: out: Gaussian augmented output. Values

r"""NASNetALarge model architecture from the `"NASNet" <https://arxiv.org/abs/1707.07012>`_ paper.

Tensorflow version of np.repeat for 2D

https://github.com/shaoanlu/experiment-with-res2net/blob/master/res2net_block.py Arguments: x: input tensor f: number

Constructs a ResNet-101 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-152 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-18 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-34 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Constructs a ResNet-50 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

Reference implementation for batch_map_coordinates

Batch version of th_map_coordinates Only supports 2D feature maps Parameters ---------- input : tf.Tensor. shape = (b, s, s)

Tensorflow verion of scipy.ndimage.map_coordinates Note that coords is transposed and only 2D is supported Parameters ---------- i

Torch version of np.repeat for 1D

unfreeze modules

Construct Xception.