Reverses the imagenet normalization applied to the input. Args: x (torch.Tensor - shape(N,3,H,W)): input tensor Returns: tor

DINOv2 ViT-B/14 model (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-B/14 model with registers (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-g/14 model (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-g/14 model with registers (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-L/14 model (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-L/14 model with registers (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-S/14 model (optionally) pretrained on the LVD-142M dataset.

DINOv2 ViT-S/14 model with registers (optionally) pretrained on the LVD-142M dataset.

Enable distributed mode Args: set_cuda_current_device: If True, call torch.cuda.set_device() to set the current PyTorch CUDA

Exponential attractor: dc = exp(-alpha*|dx|^gamma) * dx , where dx = a - c, a = attractor point, c = bin center, dc = shift in bin centermmary for exp

Fix random seeds.

Compute the results on all values of `self.nb_knn` neighbors from the full `self.max_k`

Args: student_output (BxD): backbone output of student

Cross-entropy between softmax outputs of the teacher and student networks. student_patch_tokens: (B, N, D) tensor teacher_pat

Cross-entropy between softmax outputs of the teacher and student networks.

Forward pass. Args: x (tensor): input Returns: tensor: output

Forward pass. Returns: tensor: output

Forward pass. Args: x (tensor): input Returns: tensor: output

Forward pass. Returns: tensor: output

Args: x (torch.Tensor): Input image tensor of shape (B, C, H, W) return_final_centers (bool, optional): Whether to re

Forward pass Args: x (torch.Tensor - NCHW): Input feature tensor Returns: torch.Tensor - SNE: Transformer ou

Compute log binomial distribution for x Args: x (torch.Tensor - NCHW): probabilities t (float, torch.Tensor - NCHW, o

Forward pass Args: x (torch.Tensor - NCHW): Main feature cond (torch.Tensor - NCHW): condition feature Retur

Returns tensor of bin_width vectors (centers). One vector b for every pixel

Returns tensor of bin_width vectors (centers). One vector b for every pixel

x : feature block; shape - n, c, h, w b_prev : previous bin widths normed; shape - n, prev_nbins, h, w

Args: x (torch.Tensor) : feature block; shape - n, c, h, w b_prev (torch.Tensor) : previous bin centers normed; shape

Args: x (torch.Tensor) : feature block; shape - n, c, h, w b_prev (torch.Tensor) : previous bin centers normed; shape

Args: x (torch.Tensor): Input image tensor of shape (B, C, H, W). Assumes all images are from the same domain. return

Freezes all the parameters of all the ModuleDicts children

Freezes all the parameters of all the ModuleDicts children that are exclusively used for the given bin configuration

Returns: str: The latest checkpoint file in target directory.

Returns: The rank of the current process within the local (per-machine) process group.

Returns: The size of the per-machine process group, i.e. the number of processes per machine.

Learning rate configuration for different layers of the model Args: lr (float) : Base learning rate Returns: