A scalar meter uses a deque to track a series of scaler values with a given window size. It supports calculating the median and average value

Fuses the information from the Fast pathway to the Slow pathway. Given the tensors from Slow pathway and Fast pathway, fuse information from

Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks).

Internal implementation of Pascal object detection metrics.

Image to Patch Embedding

Swish activation function: x * sigmoid(x).

Measure training stats.

Measures validation stats.

Apply for video.

Tensorflow like 'SAME' convolution wrapper for 2D convolutions

Feature Hook Helper This module helps with the setup and extraction of hooks for extracting features from internal nodes in a model by node n

A class used to get weights and activations from specified layers from a Pytorch model.

Model Exponential Moving Average (DEPRECATED) Keep a moving average of everything in the model state_dict (parameters and buffers). This

This class defines multigrid training schedule and update cfg accordingly.

Builds Non-local Neural Networks as a generic family of building blocks for capturing long-range dependencies. Non-local Network computes

Image to Patch Embedding

Image to Patch Embedding

Randomly selects a rectangle region in an image and erases its pixels. 'Random Erasing Data Augmentation' by Zhong et al. See https://

Crop the given PIL Image to random size and aspect ratio with random interpolation. A crop of random size (default: of 0.08 to 1.0) of the origin

Residual block.

Squeeze-and-Excitation (SE) block w/ Swish: AvgPool, FC, Swish, FC, Sigmoid.

Extend Sampler to support "short cycle" sampling. See paper "A Multigrid Method for Efficiently Training Video Models", Wu et al., 2019 (

Perform the multi-view ensemble for testing: each video with an unique index will be sampled with multiple clips, and the predictions of the

Measure training stats.

Vision Transformere

Vision Transformere

Vision Transformere

Vision Transformere

Basic transformation: Tx3x3, 1x3x3, where T is the size of temporal kernel.

Bottleneck transformation: Tx1x1, 1x3x3, 1x1x1, where T is the size of temporal kernel.

Box collection. BoxList represents a list of bounding boxes as numpy array, where each bounding box is represented as a row of 4 numbers, [y_mi

Naming conventions for BoxLists. Attributes: boxes: bounding box coordinates. classes: classes per bounding box. scores: scores per bou

Convenience wrapper for BoxList with masks. BoxMaskList extends the np_box_list.BoxList to contain masks as well. In particular, its constructor

Extract center crop at the same location for a list of images Args: size (sequence or int): Desired output size for the crop in format

Randomly change the brightness, contrast and saturation and hue of the clip Args: brightness (float): How much to jitter brightness. brightn

Composes several transforms Args: transforms (list of ``Transform`` objects): list of transforms to compose

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip i

Interface for object detection evalution classes. Example usage of the Evaluator: ------------------------------ evaluator = DetectionEvaluator

Naming conventions for storing the output of the detector. Attributes: source_id: source of the original image. key: unique key correspondi

Feature extractor with OrderedDict return Wrap a model and extract features as specified by the out indices, the network is partially re-buil

FeatureHookNet Wrap a model and extract features specified by the out indices using forward/forward-pre hooks. If `no_rewrite` is True, featu

Feature extractor with list return See docstring for FeatureDictNet above, this class exists only to appease Torchscript typing constraints.

GroupGather performs all gather on each of the local process/ GPU groups.

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip i

Names for the input tensors. Holds the standard data field names to use for identifying input tensors. This should be used by the decoder to iden

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip is

Model Exponential Moving Average V2 Keep a moving average of everything in the model state_dict (parameters and buffers). V2 of this modu

Normalize a clip with mean and standard deviation. Given mean: ``(M1,...,Mn)`` and std: ``(S1,..,Sn)`` for ``n`` channels, this transform wi

A class to evaluate detections.

A class to evaluate detections using Open Images V2 metrics. Open Images V2 introduce group_of type of bounding boxes and this metric handles

A class to evaluate detections using PASCAL metrics.

A class to evaluate instance masks using PASCAL metrics.

Evaluate detection result of a single image.

Extract random crop at the same location for a list of images Args: size (sequence or int): Desired output size for the crop in format

Horizontally flip the list of given images randomly with a probability 0.5

Resizes a list of (H x W x C) numpy.ndarray to the final size The larger the original image is, the more times it takes to interpolate

Rotate entire clip randomly by a random angle within given bounds Args: degrees (sequence or int): Range of degrees to select from

ResNet model builder. It builds a ResNet like network backbone without lateral connection (C2D, I3D, Slow). Christoph Feichtenhofer, Hao

ResNe(X)t 3D head. This layer performs a fully-connected projection during training, when the input size is 1x1x1. It performs a convolut

ResNe(X)t 3D stem module. Performs spatiotemporal Convolution, BN, and Relu following by a spatiotemporal pooling.

Stage of 3D ResNet. It expects to have one or more tensors as input for single pathway (C2D, I3D, Slow), and multi-pathway (SlowFast) cas

Resizes a list of (H x W x C) numpy.ndarray to the final size The larger the original image is, the more times it takes to interpolate

SlowFast model builder for SlowFast network. Christoph Feichtenhofer, Haoqi Fan, Jitendra Malik, and Kaiming He. "SlowFast networks for v

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip i

Enum class for sort order. Attributes: ascend: ascend order. descend: descend order.

Something-Something v2 (SSV2) video loader. Construct the SSV2 video loader, then sample clips from the videos. For training and validation,

The standard BN layer computes stats across all examples in a GPU. In some cases it is desirable to compute stats across only a subset of exa

Swish activation function: x * sigmoid(x).

Helper class to log information to Tensorboard.

TF-example proto feature names for object detection. Holds the standard feature names to load from an Example proto for object detection. Attr

Extract random crop at the same location for a list of images Args: size (sequence or int): Desired output size for the crop in format

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip i

Kinetics video loader. Construct the Kinetics video loader, then sample clips from the videos. For training and validation, a single clip is

Video 3D stem module. Provides stem operations of Conv, BN, ReLU, MaxPool on input data tensor for one or multiple pathways.

A class to evaluate detections using weighted PASCAL metrics. Weighted PASCAL metrics computes the mean average precision as the average precisio

A class to evaluate instance masks using weighted PASCAL metrics. Weighted PASCAL metrics computes the mean average precision as the average prec

X3D model builder. It builds a X3D network backbone, which is a ResNet. Christoph Feichtenhofer. "X3D: Expanding Architectures for Effic

X3D head. This layer performs a fully-connected projection during training, when the input size is 1x1x1. It performs a convolutional pro

X3D's 3D stem module. Performs a spatial followed by a depthwise temporal Convolution, BN, and Relu following by a spatiotemporal poo