Trans image and sequence of frames to tensor. Args: value (np.ndarray | list[np.ndarray] | Tuple[np.ndarray]): The original i

Average of key in results(list[dict]). Args: results (list[dict]): A list of dict containing the necessary data. key (str): The k

Convert a BGR image to YCbCr image. The bgr version of rgb2ycbcr. It implements the ITU-R BT.601 conversion for standard-definition telev

Generate a bivariate isotropic or anisotropic Gaussian kernel. In the isotropic mode, only `sig_x` is used. `sig_y` and `theta` is ignored. Ar

Generate a bivariate generalized Gaussian kernel. Described in `Parameter Estimation For Multivariate Generalized Gaussian Distributio

Generate a plateau-like anisotropic kernel. 1 / (1+x^(beta)) Ref: https://stats.stackexchange.com/questions/203629/is-there-a-plateau-shaped-d

Return intersection-over-union (Jaccard index) of boxes. Both sets of boxes are expected to be in (x1, y1, x2, y2) format. Arguments:

Judge whether the input value can be converted to image tensor via :func:`images_to_tensor` function. Args: value (any): The input va

Copying data to the target device. Args: data (dict): Data returned by ``DataLoader``. Returns: CollatedResu

Charbonnier loss. Args: pred (Tensor): Prediction Tensor with shape (n, c, h, w). target ([type]): Target Tensor with shape (n, c

Returns the biggest detection box and mask of a YOLO Results set

Compute optical flow using SPyNet for feature alignment. Args: lqs (tensor): Input low quality (LQ) sequence with

Compute flow from ref to supp. Note that in this function, the images are already resized to a multiple of 32. Args:

Compute the metrics from processed results. Args: results (list): The processed results of each batch. Returns:

Compute the metrics from processed results. Args: results (list): The processed results of each batch. Returns:

Compute zero padding tuple. Args: kernel_size (tuple[int]): The size of the kernel. Returns: tuple: Padding

pixel_to_class_mapping is a dict providing a mapping from pixel value to class id. e.g. if you only have a single class with id 0 and binary

Add predictions and destructed inputs (if passed) to data samples. Args: predictions (DataSample): The predictions of the model.

Fake part of D step. Args: batch_outputs (Tensor): Output of generator. batch_gt_data (Tuple[Tensor]): Batch GT data.

Real part of D step. Args: batch_outputs (Tensor): Batch output of generator. batch_gt_data (Tuple[Tensor]): Batch GT

D step with optim of GAN: Calculate losses of discriminator and run optim. Args: batch_outputs (Tensor): Batch output of

Disc Shift loss. This loss is proposed in PGGAN as an auxiliary loss for discriminator. Args: pred (Tensor): Input tensor. Retu

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). This is the same as the DropConnect impl I created for Ef

This is the function to perform every n iterations. Args: runner (Runner): runner used to drive the whole pipeline n

extract gt data from data samples. Args: data_samples (list): List of DataSample. Returns: Tensor: Extract g

Convert label of various python types to :obj:`mmengine.LabelData`. Supported types are: :class:`numpy.ndarray`, :class:`torch.Tensor`, :clas

Forward inference. Returns predictions of validation, testing, and simple inference. Args: inputs (torch.Tensor): batch i

Forward Train. Run forward of generator with ``return_lqs=True`` Args: batch_inputs (Tensor): Batch inputs.

Forward training. Returns dict of losses of training. Args: inputs (torch.Tensor): batch input tensor collated by

Fuse aesthetic and technical results into final scores

G step of GAN: Calculate losses of generator. Args: batch_outputs (Tensor): Batch output of generator. batch_gt_data

G step of GAN: Calculate losses of generator. Args: batch_outputs (Tensor): Batch output of generator. batch_gt_data

G step with optim of GAN: Calculate losses of generator and run optim. Args: batch_outputs (Tensor): Batch output of gene