Forward pass for PatchEmbed. Args: x: A float tensor of shape (b, c, h, w). Returns: A float tensor

Forward pass for Attention. Args: x: A float tensor of shape (b, n, c), where n is the sequence length and c is

Forward pass for MultiHeadSelfAttention. Args: x: A float tensor of shape (b, n, c). size: An optional (h, w

Forward pass for OverlappedSpatialReductionAttention. Args: x: A float tensor of shape (b, n, c) where n = h * w.

Forward pass for OSRABlock. Args: x: A float tensor of shape (b, n, c). size: A tuple (h, w) giving the heig

Forward pass for PatchMerging. Args: x: A float tensor of shape (b, h, w, c). Returns: A float tens

Forward pass for LePEAttention. Splits Q/K/V according to cross-shaped windows, computes attention, and returns the combined

Forward pass for CrossShapedWindowAttention. Args: x: A float tensor of shape (b, n, c), where n = h * w. si

Load pretrained parameters onto the model Args: path_or_url: the path or URL to the model parameters (checkpoint) **k

Load pretrained parameters onto the model Args: path_or_url: the path or URL to the model parameters (checkpoint) **k

Load pretrained parameters onto the model Args: path_or_url: the path or URL to the model parameters (checkpoint) **k

Get the maximum shape ratio of a contour. Args: contour: the contour from cv2.findContour Returns: the maximum shape ratio

Invert the colors of an image Args: img : torch.Tensor, the image to invert min_val : minimum value of the random shift Retu

LinkNet as described in `"LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation" <https://arxiv.org/pdf/1707.03718.pdf>`

LinkNet as described in `"LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation" <https://arxiv.org/pdf/1707.03718.pdf>`

LinkNet as described in `"LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation" <https://arxiv.org/pdf/1707.03718.pdf>`

Resnet31 architecture with Multi-Aspect Global Context Attention as described in `"MASTER: Multi-Aspect Non-local Network for Scene Text Recogniti

MASTER as described in paper: <https://arxiv.org/pdf/1910.02562.pdf>`_. >>> import torch >>> from doctr.models import master >>> model =

MobileNetV3-Large architecture as described in `"Searching for MobileNetV3", <https://arxiv.org/pdf/1905.02244.pdf>`_. >>> import torch

MobileNetV3-Large architecture as described in `"Searching for MobileNetV3", <https://arxiv.org/pdf/1905.02244.pdf>`_, with rectangular poolin

MobileNetV3-Small architecture as described in `"Searching for MobileNetV3", <https://arxiv.org/pdf/1905.02244.pdf>`_. >>> import torch

MobileNetV3-Small architecture as described in `"Searching for MobileNetV3", <https://arxiv.org/pdf/1905.02244.pdf>`_, with rectangular poolin

Perform non-max suppression, borrowed from <https://github.com/rbgirshick/fast-rcnn>`_. Args: boxes: np array of straight boxes: (*, 5),

PARSeq architecture from `"Scene Text Recognition with Permuted Autoregressive Sequence Models" <https://arxiv.org/pdf/2207.06966>`_. >>> imp

Runs docTR KIE model to analyze the input image

Runs docTR OCR model to analyze the input image

Convert a polygon to a bounding box Args: polygon: a polygon Returns: a bounding box

Converts multiclass target to relative coordinates. Args: img: Image target: tuple of target polygons and their classes names

ResNet-18 architecture as described in `"Deep Residual Learning for Image Recognition", <https://arxiv.org/pdf/1512.03385.pdf>`_. >>> import

Resnet31 architecture with rectangular pooling windows as described in `"Show, Attend and Read:A Simple and Strong Baseline for Irregular Text Rec

ResNet-34 architecture as described in `"Deep Residual Learning for Image Recognition", <https://arxiv.org/pdf/1512.03385.pdf>`_. >>> import

ResNet-34 architecture as described in `"Deep Residual Learning for Image Recognition", <https://arxiv.org/pdf/1512.03385.pdf>`_ with twice as man

ResNet-50 architecture as described in `"Deep Residual Learning for Image Recognition", <https://arxiv.org/pdf/1512.03385.pdf>`_. >>> import

Returns (return-code, stdout, stderr)

SAR with a resnet-31 feature extractor as described in `"Show, Attend and Read:A Simple and Strong Baseline for Irregular Text Recognition" <https