Convert keys in detectron pretrained ResNet models to pytorch style.

Convert keys in pycls pretrained RegNet models to mmdet style.

Copy gradients from fp16 model to fp32 weight copy.

Copy updated params from fp32 weight copy to fp16 model.

Decodes a list of string notations to specify blocks inside the network. :param string_list: a list of strings, each string is a not

Transform outputs for a single batch item into raw bbox predictions. Args: tl_heat (Tensor): Top-left corner heatmap for current

Apply deltas to shift/scale base boxes. Typically the rois are anchor or proposed bounding boxes and the deltas are network outputs used to s

r"""`Implementation of Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression, https://arxiv.org/abs/1911.08287`_. Code is

r"""Distribution Focal Loss (DFL) is from `Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

Drop connect.

Creates a efficientnet model.

Evaluation in COCO protocol. Args: results (list[list | tuple]): Testing results of the dataset. metric (str | list[s

TODO: Add expand for polygon

Trace and export a model to onnx format. Modified from https://github.com/facebookresearch/detectron2/ Args: model (nn.Module):

Extract features from images.

Returns output of the final convolution layer

Format the results to txt (standard format for Cityscapes evaluation). Args: results (list): Testing results of the datas

Args: imgs (List[Tensor]): the outer list indicates test-time augmentations and inner Tensor should have a shape

Args: img (list[Tensor]): List of tensors of shape (1, C, H, W). Typically these should be mean centered and std

During inference, the functionary of batch norm layers is turned off but only the mean and var alone channels are used, which exposes the chance t

Compute location targets for guided anchoring. Each feature map is divided into positive, negative and ignore regions. - positive reg

Compute guided anchoring targets. Args: approx_list (list[list]): Multi level approxs of each image. inside_flag_list

Generate 2D gaussian kernel. Args: radius (int): Radius of gaussian kernel. sigma (int): Sigma of gaussian function. Default: 1.

`Focal Loss <https://arxiv.org/abs/1708.02002>`_ for targets in gaussian distribution. Args: pred (torch.Tensor): The prediction.

r"""Generate 2D gaussian radius. This function is modified from the `official github repo <https://github.com/princeton-vl/CornerNet-Lite/blo

Generate base anchors. Returns: list(torch.Tensor): Base anchors of a feature grid in multiple \ feature levels.

Generates per block width from RegNet parameters. Args: initial_width ([int]): Initial width of the backbone width_sl

Transform network output for a batch into bbox predictions.

Get class names of current dataset. Args: classes (Sequence[str] | str | None): If classes is None, use default C

Get det results and gt information of a certain class. Args: det_results (list[list]): Same as `eval_map()`. annotations (list[di

Get gt priorities according to their areas. Smaller gt has higher priority. Args: gt_bboxes (Tensor): Ground truth boxes

Get the mask scores. mask_score = bbox_score * mask_iou

Get the block args and global params for a given model

Get ``num_points`` most uncertain points during test. Args: mask_pred (Tensor): A tensor of shape (num_rois, num_classes,

Get ``num_points`` most uncertain points with random points during train. Sample points in [0, 1] x [0, 1] coordinate space based on

Get sampled approxs and inside flags according to feature map sizes. Args: featmap_sizes (list[tuple]): Multi-level feature map s

Gets widths/stage_blocks of network at each stage. Args: widths (list[int]): Width in each stage. Returns: t

Compute regression and classification targets for anchors in multiple images. Args: anchor_list (list[list[Tensor]]): Mul

Get targets for ATSS head. This method is almost the same as `AnchorHead.get_targets()`. Besides returning the targets as the parent

Compute regression, classification and centerss targets for points in multiple images. Args: points (list[Tensor]): Point

Generate corner targets. Including corner heatmap, corner offset. Optional: corner embedding, corner guiding shift, centripetal shif

Compute target of mask IoU. Mask IoU target is the IoU of the predicted mask (inside a bbox) and the gt mask of corresponding gt mask

r"""`Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression <https://arxiv.org/abs/1902.09630>`_. Args:

Initialize assigner and sampler.

Initialize ``bbox_head``

Initialize ``mask_head``

Initialize ``point_head``

Initialize the weights.

Initialize the weights in detector. Args: pretrained (str, optional): Path to pre-trained weights. Defaults to No

Initialize the weights in detector. Args: pretrained (str, optional): Path to pre-trained weights. Defaults to No

Initialize the weights in detector. Args: pretrained (str, optional): Path to pre-trained weights. Defaults to No

IoU loss. Computing the IoU loss between a set of predicted bboxes and target bboxes. The loss is calculated as negative log of IoU. Arg

L1 loss. Args: pred (torch.Tensor): The prediction. target (torch.Tensor): The learning target of the prediction. Returns:

Compute deltas of proposals w.r.t. gt in the MMDet V1.x manner. We usually compute the deltas of x, y, w, h of proposals w.r.t ground truth b

Apply deltas to shift/scale base boxes in the MMDet V1.x manner. Typically the rois are anchor or proposed bounding boxes and the deltas are

Load annotation from annotation file.

Load annotation from COCO style annotation file. Args: ann_file (str): Path of annotation file. Returns: lis

Loads pretrained weights, and downloads if loading for the first time.

Load proposal from proposal file.

Compute losses of the head.