PyTorch: Loading Data from AIStore

Listing and loading data from AIS buckets (buckets that are not 3rd party backend-based) and remote cloud buckets (3rd party backend-based cloud buckets) using AISFileLister and AISFileLoader.

AIStore (AIS for short) fully supports Amazon S3, Google Cloud, and Microsoft Azure backends, providing a unified namespace across multiple connected backends and/or other AIS clusters, and more.

In the following example, we use the Caltech-256 Object Category Dataset containing 256 object categories and a total of 30607 images stored on an AIS bucket and the Microsoft COCO Dataset which has 330K images with over 200K labels of more than 1.5 million object instances across 80 object categories stored on Google Cloud.

# Imports
import os
from IPython.display import Image

from torchdata.datapipes.iter import AISFileLister, AISFileLoader, Mapper

Running the AIStore Cluster

Getting started with AIS will take only a few minutes (prerequisites boil down to having a Linux with a disk) and can be done either by running a prebuilt all-in-one docker image or directly from the open-source.

To keep this example simple, we will be running a minimal standalone docker deployment of AIStore.

# Running the AIStore cluster in a container on port 51080
# Note: The mounted path should have enough space to load the dataset

! docker run -d \
    -p 51080:51080 \
    -v <path_to_gcp_config>.json:/credentials/gcp.json \
    -e GOOGLE_APPLICATION_CREDENTIALS="/credentials/gcp.json" \
    -e AWS_DEFAULT_REGION="us-east-2" \
    -e AIS_BACKEND_PROVIDERS="gcp aws" \
    -v /disk0:/ais/disk0 \

To create and put objects (dataset) in the bucket, I am going to be using AIS CLI. But we can also use the Python SDK for the same.

! ais config cli set cluster.url=http://localhost:51080

# create bucket using AIS CLI
! ais create caltech256

# put the downloaded dataset in the created AIS bucket
! ais put -r -y <path_to_dataset> ais://caltech256/

“ais://caltech256” created (see
Files to upload:
1 3.06KiB
.jpg 30607 1.08GiB
TOTAL 30608 1.08GiB
PUT 30608 objects to “ais://caltech256”

Preloaded dataset

The following assumes that AIS cluster is running and one of its buckets contains Caltech-256 dataset.

# list of prefixes which contain data
image_prefix = ['ais://caltech256/']

# Listing all files starting with these prefixes on AIStore 
dp_urls = AISFileLister(url="http://localhost:51080", source_datapipe=image_prefix)

# list first 5 obj urls


# loading data using AISFileLoader
dp_files = AISFileLoader(url="http://localhost:51080", source_datapipe=dp_urls)

# check the first obj
url, img = next(iter(dp_files))

print(f"image url: {url}")

# view the image
# Image(

image url: ais://caltech256/002.american-flag/002_0001.jpg

def collate_sample(data):
    path, image = data
    dir = os.path.split(os.path.dirname(path))[1]
    label_str, cls = dir.split(".")
    return {"path": path, "image": image, "label": int(label_str), "cls": cls}
# passing it further down the pipeline
for _sample in Mapper(dp_files, collate_sample):

Remote cloud buckets

AIStore supports multiple remote backends. With AIS, accessing cloud buckets doesn't require any additional setup assuming, of course, that you have the corresponding credentials (to access cloud buckets).

For the following example, AIStore must be built with --gcp build tag.

--gcp, --aws, and a number of other build tags is the mechanism we use to include optional libraries in the build.

# list of prefixes which contain data
gcp_prefix = ['gcp://webdataset-testing/']

# Listing all files starting with these prefixes on AIStore 
gcp_urls = AISFileLister(url="http://localhost:51080", source_datapipe=gcp_prefix)

# list first 5 obj urls


dp_files = AISFileLoader(url="http://localhost:51080", source_datapipe=gcp_urls)
for url, file in dp_files.load_from_tar():