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

Inverse of ``locate()``. Args: t: any object with ``__module__`` and ``__qualname__``

Build a batched dataloader. The main differences from `torch.utils.data.DataLoader` are: 1. support aspect ratio grouping options 2. use

Flatten an object so it can be used for PyTorch tracing. Also returns how to rebuild the original object from the flattened outputs. Ret

Create a list of blocks of the same type that forms one ResNet stage. Args: block_class (type): a subclass of CNNBlockBa

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

Return the base directory for configurations

RoiAlign with scale 1.0.

Equivalent of ``AugmentationList(augmentations)(self)``

In-place override contents of cfg. Args: cfg: an omegaconf config object overrides: list of strings in the f

Complete Intersection over Union Loss (Zhaohui Zheng et. al) https://arxiv.org/abs/1911.08287 Args: boxes1, boxes2 (Tensor): box

Distance Intersection over Union Loss (Zhaohui Zheng et. al) https://arxiv.org/abs/1911.08287 Args: boxes1, boxes2 (Tensor): box

Args: binary_mask (ndarray): numpy array of shape (H, W), where H is the image height and W is the image width. E

Fake segmentation loss used when no suitable ground truth data was found in a batch. The loss has a value 0 and is primarily used to

Fix the __qualname__ of module members to be their exported api name, so when they are referenced in docs, sphinx can find them. Reference:

Args: tensor (bool): if True, returns 3xHxW tensor. else, returns a HxWx3 numpy array. Returns: an image, in BGR

Run model on the data_loader and evaluate the metrics with evaluator. Also benchmark the inference speed of `model.__call__` accurately.

Locate and return an object ``x`` using an input string ``{x.__module__}.{x.__qualname__}``, such as "module.submodule.class_name". Rais

A 'as_tuple=True' version of torch.nonzero to support torchscript. because of https://github.com/pytorch/pytorch/issues/38718

Patch the builtin len() function of a few detectron2 modules to use __len__ instead, because __len__ does not convert values to integers

Print main metrics in a format similar to Detectron, so that they are easy to copypaste into a spreadsheet. Args: results (Order

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

Registers a converter for the specified type. Can be used as a decorator (if converter is None), or called as a method. Args

Set the random seed for the RNG in torch, numpy and python. Args: seed (int): if None, will use a strong random seed.

Get squared Euclidean Distance Matrix Computes pairwise squared Euclidean distances between points Args: pts1: Tensor [M x D], M

Args: results (OrderedDict[dict]): task_name -> {metric -> score} Returns: bool: whether the verification succeeds or not

Args: x: (B, H, W, C) window_size (int): window size Returns: windows: (num_windows*B, window_size, window_size, C)

If one size (or aspect ratio) is specified and there are multiple feature maps, we "broadcast" anchors of that single size (or aspect ratio)

Depending on the brightness_factor, gives a lighter or darker color i.e. a color with less or more saturation than the original color

coords should be a N * 2 array-like, containing N couples of (x, y) points

img should be a numpy array, formatted as Height * Width * Nchannels

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

A data loader is created in a way similar to that of Detectron2. The main differences are: - it allows to combine datasets with differen

Build a proposal generator from `cfg.MODEL.PROPOSAL_GENERATOR.NAME`. The name can be "PrecomputedProposals" to use no proposal generator.

Return the number of samples logged in the past `window_size` iterations.

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

Crop each bitmask by the given box, and resize results to (mask_size, mask_size). This can be used to prepare training targets for Ma

Returns the config for a default multi-step LR scheduler such as "1x", "3x", commonly referred to in papers, where every 1x has the total len

Dump IR of a TracedModule/ScriptModule/Function in various format (code, graph, inlined graph). Useful for debugging. Args: mode

Returns: Boxes: tight bounding boxes around bitmasks. If a mask is empty, it's bounding box will be all zero.

Like model_zoo.get, but do not load any weights (even pretrained)

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

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

Performs non-maximum suppression (NMS) on the rotated boxes according to their intersection-over-union (IoU). Rotated NMS iteratively re

Normalize N D-dimensional embedding vectors arranged in a tensor [N, D] Args: embeddings (tensor [N, D]): N D-dimensional embedding

Paste a set of masks that are of a fixed resolution (e.g., 28 x 28) into an image. The location, height, and width for pasting each mask is d

Transpose/reshape a tensor from (N, (Ai x K), H, W) to (N, (HxWxAi), K)

Args: polygons (list[ndarray]): each array has shape (Nx2,) height, width (int) Returns: ndarray: a bool mask of sha

Args: predictions: return values of :meth:`forward()`. proposals (list[Instances]): proposals that match the features

Put multiple scalars from keyword arguments. Examples: storage.put_scalars(loss=my_loss, accuracy=my_accuracy, smoothin

Args: rgb (bool): whether to return RGB colors or BGR colors. maximum (int): either 255 or 1 Returns: ndarray: a vec

Preparation for a new round of evaluation. Should be called before starting a round of evaluation.

Args: img: same as in __init__

Makes a function retry itself after encountering pytorch's CUDA OOM error. It will first retry after calling `torch.cuda.empty_cache()`.

Return semantic segmentation predictions in the original resolution. The input images are often resized when entering semantic segmentor. Mo

Hooks are stateless by default, but can be made checkpointable by implementing `state_dict` and `load_state_dict`.

Compute and display summary metrics for evaluation results given a custom value for max_dets_per_image

Args: img_size (int): Input image size. patch_size (int): Patch size. in_chans (int): Number of input ima

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

Args: op (Callable): operation to be applied to the image, which takes in an ndarray and returns an ndarray.

Init. Args: scale_factor (float): scaling mode (str): interpolation mode

Execute inference on a list of inputs, using batch size = self.batch_size, instead of the length of the list. Inputs & outpu

Collect all configuration files (i.e. densepose_*.yaml) directly in the specified directory