A utility function to find set of an element i (uses path compression technique)

loss function (ref: F3Net-AAAI-2020)

Apply element-wise weight and reduce loss. Args: loss (Tensor): Element-wise loss. weight (Tensor): Element-wise weights.

Broadcast slot features to a 2D grid and collapse slot dimension.

1x1 convolution

3x3 convolution with padding

Builds the Slot Attention module. Args: iters: Number of iterations. num_slots: Number of slots. encoder_d

Unstack batch dimension and split into channels and alpha mask.

Args: x: (B, H, W, C) window_size (int): window size Returns: windows: (num_windows*B, window_size, window_size, C)

A function that does union of two sets of x and y (uses union by rank)

For calibrating misalignment gradient via cliping gradient technique :param optimizer: :param grad_clip: :return:

Reduce loss as specified. Args: loss (Tensor): Elementwise loss tensor. reduction (str): Options are "none", "mean" and "sum"

Constructs a Res2Net-50_v1b_26w_4s lib. Args: pretrained (bool): If True, returns a lib pre-trained on ImageNet

Broadcast slot features to a 2D grid and collapse slot dimension.

Usage: formatting `tif/tiff` files to `jpg/png` files :param _src_path: :param _dst_path: :return: