Evaluate the model performance of the whole dataset after processing all batches. Args: size (int): Length of the entire

Full connected layer.

Concat multi-level feature maps by image. [feature_level0, feature_level1...] -> [feature_image0, feature_image1...] Convert the shape of eac

Randomly disable the transform.

Accumulate per image evaluation results and store the result in self.eval. :param p: input params for evaluation :return: Non

Calculate the ious between each bbox of bboxes1 and bboxes2. Args: bboxes1 (ndarray): Shape (n, 4) bboxes2 (ndarray): Shape (k, 4

Recursively convert Tensor in inputs from ``src_type`` to ``dst_type``. Args: inputs: Inputs that to be casted. src_type (torch.d

Generate grid anchors in multiple feature levels. Args: featmap_sizes (list[tuple]): List of feature map sizes in

nn.Module: the normalization layer named "norm1"

Randomly negative the value.

Reload ``unbind`` from self.tensor.

Clip boxes according to the image shape in-place. Args: img_shape (Tuple[int, int]): A tuple of image height and width.

get both box type name and class. Args: box_type (str or type): Single box type name or class. Returns: Tuple[str, type]: A

Get annotation by index. Args: idx (int): Global index of ``ConcatDataset``. Returns: dict: The idx-th annot

Get original model if the input model is a model wrapper. Args: model (nn.Module): A model may be a model wrapper. Returns:

Check if is ResNet building block.

Adjust the predicted bboxes from bbox head. Args: grid_preds (dict[str, Tensor]): dictionary outputted by forward

Set ``pred_label``.

Convert value to torch.Tensor.

bbox post-processing method. The boxes would be rescaled to the original image scale and do the nms operation. Usually with_nms is Fa

Extra repr string when auto-generating magnitude is enabled.

Get palette from various inputs. Args: palette (list[tuple] | str | tuple): palette inputs. num_classes (int): the number of clas

Initialize a detector from config file. Args: config (str, :obj:`Path`, or :obj:`mmengine.Config`): Config file path, :obj:`P

Check if is one of the norms.

Random select some elements from the gallery. If `gallery` is a Tensor, the returned indices will be a Tensor; If `gallery` is a ndar

Randomly generate magnitude.

Set ``pred_label``.

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

Tracking forward function. Args: model (nn.Module): MOT model. img (Tensor): of shape (T, C, H, W) encoding input ima

Box head forward function used in both training and testing. Args: stage (int): The current stage in Cascade RoI Head.

Calculate accuracy according to the prediction and target. Args: pred (torch.Tensor): The model prediction, shape (N, num_class)

Convert bbox coordinates from (x1, y1, x2, y2) to (cx, cy, ratio, h). Args: bbox (Tensor): Shape (n, 4) for bboxes. Returns:

Convert boxes from source type to destination type. If ``boxes`` is a instance of BaseBoxes, the ``src_type`` will be set as the type of ``bo

Process one batch of data samples and predictions. The processed results should be stored in ``self.results``, which will be used to c

Get batch data from dataset. Args: inputs (Iterable): An iterable dataset. chunk_size (int): Equivalent to batch size

Initialize conv sequential for CornerHead.

Mask head forward function used in both training and testing. Args: stage (int): The current stage in Cascade RoI Head.

Gather tensors with different first-dimension size and concat to one tenosr. Note: Only the first dimension should be different.

Check whether the anchors are inside the border. Args: flat_anchors (torch.Tensor): Flatten anchors, shape (n, 4). valid_flags (t

r"""fetches neighbors features from a given neighbor idx Args: x (Tensor): input feature Tensor :math: `\

Concatenate boxes with type of tensor or box type. Args: data_list (List[Union[Tensor, :obj:`BaseBoxes`]]): A list of tensors

Reload ``chunk`` from self.tensor.

Filter ground truth (GT) instances by score and/or size. Args: batch_data_samples (SampleList): The Data Samples. It usually

Filter results using score threshold and topk candidates. Args: scores (Tensor): The scores, shape (num_bboxes, K). score_thr (fl

Generate grid points of multiple feature levels. Args: featmap_sizes (list[tuple]): List of feature map sizes in

List all models available in MMPretrain. Args: pattern (str | None): A wildcard pattern to match model names. Defaults to Non

Load data list.

Process one batch of data samples and predictions. The processed results should be stored in ``self.results``, which will be used to

Convert the model into training mode while keep normalization layer frozen.

Translate the BitmapMasks. Args: out_shape (tuple[int]): Shape for output mask, format (h, w). offset (int | float):

Translate boxes in-place. Args: distances (Tuple[float, float]): translate distances. The first is horizontal dis

Computes one-to-one matching based on the weighted costs. This method assign each query prediction to a ground truth or background. T

The function is to build position embedding for model to obtain the position information of the image patches. Args: patches_resoluti

Apply transformation `pred_bboxes` to `boxes`. Args: bboxes (torch.Tensor or :obj:`BaseBoxes`): Basic boxes. Shape

The unified entry for a forward process in both training and test. The method should accept three modes: "tensor", "predict" and "loss":

Inference image(s) with the detector. Args: model (nn.Module): The loaded detector. imgs (str, ndarray, Sequence[str/ndarray]):

Load result file and return a result api object. :param resFile (str) : file name of result file :return: res (obj) : r

Reload ``new_empty`` from self.tensor.

nn.Module: the normalization layer named "norm2"

Process one batch of data samples and predictions. The processed results should be stored in ``self.results``, which will be used to c

Call function to pad images, masks, semantic segmentation maps. Args: results (dict): Result dict from loading pipeline.

Generate valid flags of anchors in a feature map. Args: featmap_sizes (list(tuple[int, int])): List of feature map sizes in

Add predictions to `DetDataSample`. Args: data_samples (list[:obj:`DetDataSample`], optional): A batch of data sa

Attention function.

Apply transformation `pred_bboxes` to `boxes`. Args: bboxes (torch.Tensor or :obj:`BaseBoxes`): Basic boxes. Shape

Encode bitmap mask to RLE code. Args: mask_results (list): bitmap mask results. Returns: list | tuple: RLE encoded mask.

Get ann ids that satisfy given filter conditions. default skips that filter. :param vidIds (int array) : get anns for given vids

Get class names of a dataset.

Find boxes inside the image. Args: img_shape (Tuple[int, int]): A tuple of image height and width. all_inside (bool):

Load anns with the specified ids. :param ids (int array) : integer ids specifying anns :return: anns (object array) : loaded an

Load cats with the specified ids. :param ids (int array) : integer ids specifying cats :return: cats (object array) : loaded ca

Parse the package dependencies listed in a requirements file but strips specific versioning information. Args: fname (str): path to r

print process memory info.

Replace the string "${key}" with the corresponding value. Replace the "${key}" with the value of ori_cfg.key in the config. And support repla

Reset the buffer of the tracker.

Makes a function retry itself after encountering pytorch's CUDA OOM error. It will first retry after calling `torch.cuda.empty_cache()`. If t

Generate sparse anchors according to the ``prior_idxs``. Args: prior_idxs (Tensor): The index of corresponding anchors

Transpose and gather feature according to index. Args: feat (Tensor): Target feature map. ind (Tensor): Target coord index.