Arguments: pred_keypoint_logits (Tensor): A tensor of shape (N, K, S, S) where N is the total number of instances in the batc

Returns: dict[name -> number]: the scalars that's added in the current iteration.

Neural network layers that makes predictions from input features.

List all registered datasets. Returns: list[str]

Load precomputed object proposals into the dataset. The proposal file should be a pickled dict with the following keys: - "ids": list[i

Load Pascal VOC detection annotations to Detectron2 format. Args: dirname: Contain "Annotations", "ImageSets", "JPEGImages"

Log no more than once per n seconds. Args: lvl (int): the logging level msg (str): n (int): name (str): name

Compute the mask prediction loss defined in the Mask R-CNN paper. Args: pred_mask_logits (Tensor): A tensor of shape (B, C, Hmask, Wm

Create dataset dicts for panoptic segmentation, by merging two dicts using "file_name" field to match their entries. Args: detec

Yields: A context within which all the events added to this storage will be prefixed by the name scope.

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

Args: boxes (ndarray): an Nx5 numpy array of (x_center, y_center, width, height, angle_degrees) format

Args: masks (tensor): A tensor of shape (B, M, M) representing B masks. padding (int): Number of cells to pad on all sides.

Parse a PASCAL VOC xml file.

Paste a single mask in an image. This is a per-box implementation of :func:`paste_masks_in_image`. This function has larger quantization

Rearrange the tensor layout from the network output, i.e.: list[Tensor]: #lvl tensors of shape (N, A x K, Hi, Wi) to per-image prediction

Normalize, pad and batch the input images.

Args: dataset_dicts (list[dict]): list of dataset dicts. class_names (list[str]): list of class names (zero-indexed).

Process an input/output pair. Args: input: the input that's used to call the model. output: the return value

Rasterize the polygons into a mask image and crop the mask content in the given box. The cropped mask is resized to (mask_size, mask_size

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

Register a dataset in COCO's json annotation format for instance detection, instance segmentation and keypoint detection. (i.e., Type 1 a

Register a COCO panoptic segmentation dataset named `name`. The annotations in this registered dataset will contain both instance annotations

Register a dataset in LVIS's json annotation format for instance detection and segmentation. Args: name (str): a name that identifie

remove ops if its output is not used or not in external_output

In PyTorch nn.Linear has to take 2D tensor, this often leads to reshape a 4D tensor to 2D by calling .view(). However this (dynamic) resh

Rename the op_id-th operator in predict_net, change it's input_id-th input's name to the new_name. It also does automatic re-route and ch

Args: gt_objectness_logits (Tensor): shape (N,), each element in {-1, 0, 1} representing ground-truth objectness labels with:

Run the caffe2 model on given inputs, recording the shape and draw the graph. predict_net/init_net: caffe2 model. tensor_inputs: a list

Save the graph as SVG format. Args: output_file (str): a SVG file inputs: optional inputs given to the model

Save the model as caffe2's protobuf format. Args: output_dir (str): the output directory to save protobuf files.

Scale the box with horizontal and vertical scaling factors

Args: boxes (tensor): A tensor of shape (B, 4) representing B boxes with 4 coords representing the corners x0, y0, x1, y1,

Args: proposals (list[Instances]): a list of N Instances, where N is the number of images. Returns: proposals: o

Produce additional matches for predictions that have only low-quality matches. Specifically, for each ground-truth G find the set of

Create configs and perform basic setups.

Create configs and perform basic setups.

Load custom environment setup by importing a Python source file or a module, and run the setup function.

Perform environment setup work. The default setup is a no-op, but this function allows the user to specify a Python source file or a module in

Compute the smooth L1 loss for box regression. Returns: scalar Tensor

Compute the softmax cross entropy loss for box classification. Returns: scalar Tensor

Helper function to synchronize (barrier) among all processes when using distributed training

Args: cfg (CfgNode): model (nn.Module): evaluators (list[DatasetEvaluator] or None): if None, will call

Arguments: boxes: Nx4 tensor, the boxes to draw the keypoints to Returns: heatmaps: A tensor

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

Transform keypoint annotations of an image. Args: keypoints (list[float]): Nx3 float in Detectron2 Dataset format. transform

Update the model with precise statistics. Users can manually call this method.

Convert pred_mask_logits to estimated foreground probability masks while also extracting only the masks for the predicted classes in pred_ins

Compute the mask prediction loss defined in the Mask R-CNN paper. Args: pred_mask_logits (Tensor): A tensor of shape (B, C, Hmask, W

A function used to visualize images and proposals. It shows ground truth bounding boxes on the original image and up to 20 predicted

Compute VOC AP given precision and recall. If use_07_metric is true, uses the VOC 07 11-point method (default:False).

rec, prec, ap = voc_eval(detpath, annopath, imagesetfile,

Apply the horizontal flip transform on rotated boxes. Args: rotated_boxes (ndarray): Nx5 floating point array of (x_cent

Interface for Python