dgs.models.dataset

Modules for loading data, including datasets and data loaders. The modules are a combination of my custom BaseModule and a regular torch Dataset. Additionally, I implemented a Dataset for the PoseTrack21 dataset that can be loaded within torchreid.

Module Functions

get_dataset(name)

Given the name of one dataset, return an instance.

get_multi_dataset(config, path, ds_name[, ...])

Create a concatenated dataset from the given configuration and path.

register_dataset(name, new_ds)

Register a new dataset module in :data:DATASETS, to be able to use it from configuration files.

Modules

MOT

Datasets and helpers for the MOT datasets.

alphapose

Load bboxes and poses from existing .json file, generated by AP.

dataset

Module for handling data loading and preprocessing using torch Datasets.

keypoint_rcnn

Use keypointrcnn_resnet50_fpn() to predict the key points and bounding boxes of each image.

posetrack21

Load bboxes and poses from an existing .json file of the PoseTrack21 dataset.

torchreid_pose_dataset

Default Datasets for pose-based data.