Vision Transformer with support for patch or hybrid CNN input stage

r""" LayerNorm that supports two data formats: channels_last (default) or channels_first. The ordering of the dimensions in the inputs. channels_

Wraps COCOeval for 2D or 3D box evaluation depending on mode

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

A wrapper for ``torch.nn.MultiheadAttention``. This module implements MultiheadAttention with identity connection, and positional encoding is

Image to Patch Embedding

Transformer blocks with support of window attention and residual propagation blocks

r""" ConvNeXt Block. There are two equivalent implementations: (1) DwConv -> LayerNorm (channels_first) -> 1x1 Conv -> GELU -> 1x1 Conv; all in (N

r""" ConvNeXt A PyTorch impl of : `A ConvNet for the 2020s` - https://arxiv.org/pdf/2201.03545.pdf Args: in_chans (int)

Resolves CubeRCNN's model zoo files.

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

Implement the scaled dot product attention with softmax. Arguments --------- softmax_scale: The temperature to use for the softmax att

Very simple multi-layer perceptron (also called FFN)

Class for COCO-like dataset object. Not inherently related to use with Detectron2 or training per se.

Params for the Omni evaluation API

r""" Notes ----- Adapted from https://github.com/fundamentalvision/BEVFormer/blob/master/projects/mmdet3d_plugin/datasets/pipelines/transf

The standard bottleneck residual block without the last activation layer. It contains 3 conv layers with kernels 1x1, 3x3, 1x1.

BBox3DL1Cost. Args: weight (int | float, optional): loss_weight

r"""Feature Pyramid Network. This is an implementation of paper `Feature Pyramid Networks for Object Detection <https://arxiv.org/abs/1612.03

Computes one-to-one matching between predictions and ground truth. This class computes an assignment between the targets and the predictions b

Position embedding with learnable embedding weights. Args: num_feats (int): The feature dimension for each position along x-ax

Args: c_dim (int): dimension of latent code c f_dim (int): feature dimension

This module implements SimpleFeaturePyramid in :paper:`vitdet`. It creates pyramid features built on top of the input feature map.

Position encoding with sine and cosine functions. See `End-to-End Object Detection with Transformers <https://arxiv.org/pdf/2005.12872>`_ for

This module implements Vision Transformer (ViT) backbone in :paper:`vitdet`. "Exploring Plain Vision Transformer Backbones for Object Detecti