Vision Transformer

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

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

Vision Transformer

Convert a list of m (H x W x C) numpy.ndarrays in the range [0, 255] to a torch.FloatTensor of shape (C x m x H x W) in the range [0, 1.0]

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

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 Arg

Identity layer to replace last fully connected layer

Inference parameters that are passed to the main model in order to efficienly calculate and store the context during inference.

compute sinusoid encoding.

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

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 (h,

Vision Transformer with support for patch or hybrid CNN input stage

Vision Transformer with support for patch or hybrid CNN input stage

VisionMamba adapted for MAE Pretraining

UCF101 video loader. Construct the UCF101 video loader, then sample clips from the videos. For training and validation, a single clip is

Subclass torch's LayerNorm to handle fp16.

Scale vector by element multiplications.

Track a series of values and provide access to smoothed values over a window or the global series average.

UCF101 video loader. Construct the UCF101 video loader, then sample clips from the videos. For training and validation, a single clip is

Vision Transformer with support for patch or hybrid CNN input stage

Attentive Classifier

Attentive Pooler

This class implements a weighted sampler for distributed training. See https://pytorch.org/docs/stable/_modules/torch/utils/data/distributed.

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

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

LayerNorm that supports two data formats: channels_last (default) or channels_first. The ordering of the dimensions in the inputs. channels_last

Implement resource monitoring on a per-worker basis. The sampling occurs every monitor_interval seconds and writes the log every log_interval

Override torch dataloader to return out of order.

Load video phase recognition dataset.

Vision Transformer with support for patch or hybrid CNN input stage

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

Covid XRay dataset.

Load your own video classification dataset. Parameters ---------- root : str, required. Path to the root folder storing the datase

Vision Transformer

computes and stores the average and current value

Helper to map indices of concatenated/mixed datasets to the sample index for the corresponding dataset.

ConvNeXt V2 Args: in_chans (int): Number of input image channels. Default: 3 num_classes (int): Number of classes for classifica

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

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

Subclass torch's LayerNorm to handle fp16.

r"""MetaFormer A PyTorch impl of : `MetaFormer Baselines for Vision` - https://arxiv.org/abs/2210.13452 Args: in_chans

Image to Patch Embedding

Image to Patch Embedding

Image to Patch Embedding

Image to Patch Embedding

Image to Patch Embedding

Load video phase recognition dataset.

Load video phase recognition dataset.

Vision Transformer with support for patch or hybrid CNN input stage

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

Track a series of values and provide access to smoothed values over a window or the global series average.

Track a series of values and provide access to smoothed values over a window or the global series average.

Video classification dataset for pre-extracted frames.