r"""Inject some information about the relative or absolute position of the tokens in the sequence. The positional encodings have the same dime

This structure stores a list of bezier curves as a Nx16 torch.Tensor. It will support some common methods about bezier shapes (`area`, `c

r"""Allows the model to jointly attend to information from different representation subspaces. See reference: Attention Is All You Need ..

Intersetion Over Union (IoU) loss which supports three different IoU computations: * IoU * Linear IoU * gIoU

This module is used in FCOS to generate extra layers

This module is used in RetinaNet and FCOS to generate extra layers, P6 and P7 from C5 or P5 feature.

r"""A transformer model. User is able to modify the attributes as needed. The architecture is based on the paper "Attention Is All You Need". Ashi

Same as :class:`detectron2.modeling.ProposalNetwork`. Use one stage detector and a second stage for instance-wise prediction.

Same as :class:`DetectronCheckpointer`, but is able to convert models in AdelaiDet, such as LPF backbone.

This caller enables the default Detectron2 mapper to read an additional basis semantic label

Dataloader for Local Mask. Arguments: root (string): filepath to dataset folder. dataset (string): mask to use (eg. 'train',

r"""NaiveGroupNorm implements Group Normalization with the high-level matrix operations in PyTorch. It is a temporary solution to export GN by ONN

ROIPooler with option to assign level by max length. Used by top modules.

r"""TransformerDecoder is a stack of N decoder layers Args: decoder_layer: an instance of the TransformerDecoderLayer() class (required).

r"""TransformerDecoderLayer is made up of self-attn, multi-head-attn and feedforward network. This standard decoder layer is based on the paper "A

r"""TransformerEncoder is a stack of N encoder layers Args: encoder_layer: an instance of the TransformerEncoderLayer() class (required).

r"""TransformerEncoderLayer is made up of self-attn and feedforward network. This standard encoder layer is based on the paper "Attention Is All Y

A predictor that runs the model asynchronously, possibly on >1 GPUs. Because rendering the visualization takes considerably amount of time,

This module implements Feature Pyramid Network. It creates pyramid features built on top of some input feature maps.

Large Kernel Matters -- https://arxiv.org/abs/1703.02719

Class to calculate mean-iou using fast_hist method

Should freeze bn

To do the mask encoding of PCA. components_: (tensor), shape (n_components, n_features) if agnostic=True

Instance-aware cropping.

This module implements Feature Pyramid Network. It creates pyramid features built on top of some input feature maps.

Evaluate text proposals and recognition.

This is the same Trainer except that we rewrite the `build_train_loader`/`resume_or_load` method.

A modified version of FCOS with Bezier regression

Main class for BlendMask architectures (see https://arxiv.org/abd/1901.02446).

Main class for CondInst architectures (see https://arxiv.org/abs/2003.05664).

Deformable convolutional layer with configurable deformable groups, dilations and groups. Code is from: https://github.com/facebookr

Implement FCOS (https://arxiv.org/abs/1904.01355).

Implement MEInst (https://arxiv.org/abs/2003.11712).

Same as :class:`detectron2.modeling.ProposalNetwork`. Uses "instances" as the return key instead of using "proposal".

Same as :class:`detectron2.modeling.ProposalNetwork`. Use one stage detector and a second stage for instance-wise prediction.

TextHead performs text region alignment and recognition. It is a simplified ROIHeads, only ground truth RoIs are used during trainin

Instance-aware cropping.