Feature map interpolation.

Converts the model to be resolution-agnostic. This method replaces standard adaptive pooling layers and performs a dry run w

r"""Fills the input Tensor with values drawn from a truncated normal distribution. From: https://github.com/rwightman/pytorch-image-models/bl

Get the root logger. The logger will be initialized if it has not been initialized. By default a StreamHandler will be added. If `log_file` i

Get the root logger. The logger will be initialized if it has not been initialized. By default a StreamHandler will be added. If `log_file` i

r"""Change to Y channel Args: image tensor: tensor with shape (N, 3, H, W) in range [0, 1]. Returns: image tensor: Y channel o

Get the root logger. The logger will be initialized if it has not been initialized. By default a StreamHandler will be added. If `log_file` i

Read an image from bytes. Args: content (bytes): Image bytes got from files or other streams. flag (str): Flags specifying the co

Save networks and training state.

Save networks and training state.

Scan a directory to find the interested files. Args: dir_path (str): Path of the directory. suffix (str | tuple(str), optional):

Make lmdb from images. Contents of lmdb. The file structure is: example.lmdb ├── data.mdb ├── lock.mdb ├── meta_info.txt The

Prepare image path list and keys for DIV2K dataset. Args: folder_path (str): Folder path. Returns: list[str]: Image path lis

Scan a directory to find the interested files. Args: dir_path (str): Path of the directory. suffix (str | tuple(str), optional):

Numpy array to tensor. Args: imgs (list[ndarray] | ndarray): Input images. bgr2rgb (bool): Whether to change bgr to rgb.

Load file form http url, will download models if necessary. Ref:https://github.com/1adrianb/face-alignment/blob/master/face_alignment/utils.py

Feature map up-sampling.

Feature map down-sampling.

PyTorch version of cv2.filter2D Args: img (Tensor): (b, c, h, w) kernel (Tensor): (b, k, k)

Get bare model, especially under wrapping with DistributedDataParallel or DataParallel.

Read a sequence of images from a given folder path. Args: path (list[str] | str): List of image paths or image folder path. requi

Calculate Gram matrix. Args: x (torch.Tensor): Tensor with shape of (n, c, h, w). Returns: torch.Tensor: Gra

Get bare model, especially under wrapping with DistributedDataParallel or DataParallel.

Remove CLS token.

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

Scan a directory to find the interested files. Args: dir_path (str): Path of the directory. suffix (str | tuple(str), optional):

r"""Compute similarity_map between two tensors using Dice-like equation. Args: map_x: Tensor with map to be compared map_y: Tensor

Randomly crops the input tensor(s) to the specified size and number of crops. Args: - input_list (list or tensor): List of input ten

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Matlab like conv2, weights needs to be flipped. Args: input (tensor): (b, c, h, w) weight (tensor): (out_ch, in_ch, kh, kw), conv

Get bare model, especially under wrapping with DistributedDataParallel or DataParallel.

Numpy array to tensor. Args: imgs (list[ndarray] | ndarray): Input images. bgr2rgb (bool): Whether to change bgr to rgb.

Save networks. Args: net (nn.Module | list[nn.Module]): Network(s) to be saved. net_label (str): Network label.