WebDataset

WebDataset is a PyTorch Dataset (IterableDataset) implementation providing efficient access to datasets stored in POSIX tar archives and uses only sequential/streaming data access. This brings substantial performance advantage in many compute environments, and it is essential for very large scale training.

While WebDataset scales to very large problems, it also works well with smaller datasets and simplifies creation, management, and distribution of training data for deep learning.

WebDataset implements standard PyTorch IterableDataset interface and works with the PyTorch DataLoader. Access to datasets is as simple as:

import webdataset as wds

dataset = wds.WebDataset(url).shuffle(1000).decode("torchrgb").to_tuple("jpg;png", "json")
dataloader = torch.utils.data.DataLoader(dataset, num_workers=4, batch_size=16)

for inputs, outputs in dataloader:
    ...

In that code snippet, url can refer to a local file, a local HTTP server, a cloud storage object, an object on an object store, or even the output of arbitrary command pipelines.

WebDataset fulfills a similar function to Tensorflow's TFRecord/tf.Example classes, but it is much easier to adopt because it does not actually require any kind of data conversion: data is stored in exactly the same format inside tar files as it is on disk, and all preprocessing and data augmentation code remains unchanged.

Documentation

• Getting Started -- a simple demonstration of how to use WebDataset
• How It Works -- WebDataset is just a collection of PyTorch IterableDataset implementations
• Decoding -- WebDataset uses file-based representations internally and its Decoder makes turning these into PyTorch tensors easy
• Desktop Usage and Caching -- WebDataset is great for desktop development and supports built-in caching
• Creating Datasets in WebDataset Format -- there are a variety of tools for creating datasets in WebDataset format
• Writing WebDataset from Python -- you can write WebDataset datasets directly from Python
• Multinode Training -- what you need to know to handle distributed, multi-node training with WebDataset
• Data Sources -- accessing data on local disks, on web servers, or in cloud storage buckets

Installation

$ pip install webdataset

For the Github version:

$ pip install git+https://github.com/tmbdev/webdataset.git

Documentation: ReadTheDocs

Introductory Videos

Here are some videos talking about WebDataset and large scale deep learning:

• Introduction to Large Scale Deep Learning
• Loading Training Data with WebDataset
• Creating Datasets in WebDataset Format
• Tools for Working with Large Datasets

More Examples

• loading videos
• splitting raw videos into clips for training
• converting the Falling Things dataset

Related Libraries and Software

The AIStore server provides an efficient backend for WebDataset; it functions like a combination of web server, content distribution network, P2P network, and distributed file system. Together, AIStore and WebDataset can serve input data from rotational drives distributed across many servers at the speed of local SSDs to many GPUs, at a fraction of the cost. We can easily achieve hundreds of MBytes/s of I/O per GPU even in large, distributed training jobs.

The tarproc utilities provide command line manipulation and processing of webdatasets and other tar files, including splitting, concatenation, and xargs-like functionality.

The tensorcom library provides fast three-tiered I/O; it can be inserted between AIStore and WebDataset to permit distributed data augmentation and I/O. It is particularly useful when data augmentation requires more CPU than the GPU server has available.

You can find the full PyTorch ImageNet sample code converted to WebDataset at tmbdev/pytorch-imagenet-wds