Concatenates a list of Boxes into a single Boxes Arguments: boxes_list (list[Boxes]) Returns: Boxes

Collects all the annotations / detections for that particular image.

Efficient version of torch.cat that avoids a copy if there is only a single element in a list

Args: norm (str or callable): Returns: nn.Module or None: the normalization layer

Clone the Boxes. Returns: Boxes

Transpose/reshape a tensor from (N, (A x K), H, W) to (N, (HxWxA), K)

Returns: bool: whether the field called `name` exists.

Args: tensors: a tuple or list of `torch.Tensors`, each of shape (Hi, Wi) or (C_1, ..., C_K, Hi, Wi) where K >= 1

Returns: list[(number, iteration)]: content of the current buffer.

Args: box: can be a k-tuple, k-list or an Nxk array/tensor, where k = 4 or 5 from_mode, to_mode (BoxMode) Re

Args: top (int): number of rows of `pad_value` to add on top. left (int): number of columns of `pad_value` to add on

Get box regression transformation deltas (dx, dy, dw, dh) that can be used to transform the `src_boxes` into the `target_boxes`. That

Args: top (int): number of rows of `pad_value` to add on top. left (int): number of columns of `pad_value` to add on

Add a scalar `value` to the `HistoryBuffer` associated with `name`. Args: smoothing_hint (bool): a 'hint' on whether thi

Set attributes from the input list of parameters. Args: params (list): list of parameters.

Log only for the first n times. Args: lvl (int): the logging level msg (str): n (int): key (str or tuple[str

Resize the output instances. The input images are often resized when entering an object detector. As a result, we often need the outputs

:meth: read the object from a dict

Computes the area of all the boxes. Returns: torch.Tensor: a vector with areas of each box.

Given two lists of boxes of size N and M, compute the IoU (intersection over union) between __all__ N x M pairs of boxes. The box ord

Returns: (float): the total number of seconds since the start/reset of the timer, excluding the time when the tim

Clip (in place) the boxes by limiting x coordinates to the range [0, width] and y coordinates to the range [0, height]. Args

Returns: The :class:`EventStorage` object that's currently being used. Throws an error if no :class`EventStorage` is currently en

Create a ResNet instance from config. Returns: ResNet: a :class:`ResNet` instance.

Dump model and checkpointables to a file. Args: name (str): name of the file. kwargs (dict): extra arbitrary

r""" decode feature maps, width height, regression to detections results. Args: fmap (Tensor): input feature map.

Return intersection-over-union (Jaccard index) of boxes. Both sets of boxes are expected to be in (x_center, y_center, width, height, an

Apply transformation `deltas` (dx, dy, dw, dh) to `boxes`. Args: deltas (Tensor): transformation deltas of shape (N, k*4

Create a small table using the keys of small_dict as headers. This is only suitable for small dictionaries. Args: small_dict (di

Register the given object under the the name `obj.__name__`. Can be used as either a decorator or not. See docstring of this class fo

Apply a list of :class:`TransformGen` on the input image, and returns the transformed image and a list of transforms. We can

find a given key and its value in config Args: key (str): the string you want to find show (bool): if show i

Loads ground truth from a COCO-style annotation file. If path is not specified, this will download the COCO annotations for the year and

Set flag according to image aspect ratio. Images with aspect ratio greater than 1 will be set as group 1, otherwise group 0.

Apply the transform on an image. Args: img (ndarray): of shape NxHxWxC, or HxWxC or HxW. The array can be

Same as torchvision.ops.boxes.batched_nms, but safer.

Transpose/reshape a tensor from (N, (A x K), H, W) to (N, (HxWxA), K)

A wrapper around :function:`torch.nn.functional.grid_sample` to support 3D point_coords tensors. Unlike :function:`torch.nn.functional.grid_s

Read an image into the given format. Will apply rotation and flipping if the image has such exif information. Args: file_name (s

Smooth L1 loss defined in the Fast R-CNN paper as: | 0.5 * x ** 2 / beta if abs(x) < beta smoothl1(x) = |

Extra keyword arguments supported in addition to those in `torch.nn.Conv2d`: Args: norm (nn.Module, optional): a normali

Convert BatchNorm/SyncBatchNorm in module into FrozenBatchNorm. Args: module (torch.nn.Module): Returns:

Args: activation (EasyDict or str): Returns: nn.Module or None: the activation layer

Returns: dict: a dict which maps names (str) to data of the fields Modifying the returned dict will modify this instance

Get the shapes from a jit value object. Args: val (torch._C.Value): jit value object. Returns: list(int): return a list

Merge config (keys, values) in a list (e.g., from command line) into this CfgNode. Args: cfg_list (list): cfg_li

Find boxes that are non-empty. A box is considered empty, if either of its side is no larger than threshold. Returns:

Open a stream to a URI, similar to the built-in `open`. Args: path (str): A URI supported by this PathHandler R

Args: boxes (Boxes, RotatedBoxes or ndarray): either a :class:`Boxes`, or an Nx4 numpy array of XYXY_ABS format f

Set the field named `name` to `value`. The length of `value` must be the number of instances, and must agree with other exist

Initialize the cvpods logger and set its verbosity level to "INFO". Args: output (str): a file name or a directory to save log. If N

Performs a single optimization step. Arguments: closure (callable, optional): A closure that reevaluates the model and returns the

Run training. Returns: OrderedDict of results, if evaluation is enabled. Otherwise None.

Load and prepare dataset dicts for instance detection/segmentation and semantic segmentation. Args: dataset_name

cbl = conv + batch_norm + leaky_relu

Create an :class:`Instances` object used by the models, from instance annotations in the dataset dict. Args: annos (list[dict]):

Apply the transform on coordinates. Args: coords (ndarray): floating point array of shape Nx2. Each row is (x, y).

Raise an error if the image does not match the size specified in the dict.

Copies a source path to a destination path. Args: src_path (str): A URI supported by this PathHandler dst_pa

Args: dataset_names (list[str]): list of dataset names Returns: ndarray[int]: a vector of size=#keypoints, storing the

Filter out empty instances in an `Instances` object. Args: instances (Instances): by_box (bool): whether to filter out insta

Returns the field called `name`.

Returns: ndarray: the visualized image of shape (H, W, 3) (RGB) in uint8 type. The shape is scaled w.r.t the input

Compare the "keep" result of two nms call. They are allowed to be different in terms of edit distance due to floating point precision iss

Plots a summary model for each run in this TIDE object. Images will be outputted to out_dir, which will be created if it doesn't exis

Args: stem (nn.Module): a stem module stages (list[list[ResNetBlock]]): several (typically 4) stages,

Checks if there is a resource at the given URI. Args: path (str): A URI supported by this PathHandler Returns:

Checks if there the resource at the given URI is a file. Args: path (str): A URI supported by this PathHandler

Return the median of the latest `window_size` values in the buffer.

Given two lists of rotated boxes of size N and M, compute the IoU (intersection over union) between __all__ N x M pairs of boxes. The

Compute and display summary metrics for evaluation results. Note this functin can *only* be applied on the default parameter setting