Unable to download coco dataset with gsutil [ServiceException, AccessDeniedException]

[johannabar]

Hi there, when trying to download the coco dataset with gsutil following this page, I am having the following error:

gsutil -m rsync gs://images.cocodataset.org/val2017 val2017
ServiceException: 401 Anonymous caller does not have storage.objects.list access to images.cocodataset.org.

If I log in to gcloud, I am having the following error:

AccessDeniedException: 403 username@****.com does not have storage.objects.list access to images.cocodataset.org.

Probably related to issue #216.

Does anyone know how to fix that?

[Tanmay-Kulkarni101]

Still down

[alexxcollins]

still down - guess this isn't being resolved as its been an issue for 2 years now

[Saurabhbaghel]

still down

[BillyMazotti]

still down

[surdarla]

still down

[elsampsa]

still down.. microsoft or whoever is maintaining this, obviously, doesn't give a **it

[jacob-morosco]

still down

[LeoAKALiu]

still down

[xf15]

still 404

[xf15]

what dataset would be alternative to ms coco for object detection in scene then?

[surdarla]

There is fiftyone library that can get COCO dataset, not from gsutil see the documentation here

!pip install -q fiftyone
!pip install "opencv-python-headless<4.3" #if your using it in colab

import fiftyone as fo
import fiftyone.zoo as foz

# The path to the source files that you manually downloaded
source_dir = "/path/to/dir-files"

ds = foz.load_zoo_dataset('coco-2014', source_dir = source_dir)

ds.persistent = True

session = fo.launch_app(ds)

ds

output will be something like this

Name:        coco-2014
Media type:  image
Num samples: 164062
Persistent:  True
Tags:        []
Sample fields:
    id:           fiftyone.core.fields.ObjectIdField
    filepath:     fiftyone.core.fields.StringField
    tags:         fiftyone.core.fields.ListField(fiftyone.core.fields.StringField)
    metadata:     fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.metadata.ImageMetadata)
    ground_truth: fiftyone.core.fields.EmbeddedDocumentField(fiftyone.core.labels.Detections)

[Guardiaans]

still down..

[LjIA26]

for me too

[markp3rry]

Still down, got the BucketNotFoundException today on val2017

[rjacobs914]

I was able to download via curl in the Terminal. For example:

curl -O http://images.cocodataset.org/zips/val2017.zip

[misslibra]

still down

[mahmudhasankhan]

Why's there no fix for this? I get the same error when i try to download it with gsutil like most people here have commented about.

BucketNotFoundException: 404 gs://images.cocodataset.org bucket does not exist.

[jayaBalaR]

When I tried to use gsutil in colab and I got the exception as above

[jayaBalaR]

[somidi85]

Still down! BucketNotFoundException: 404 gs://images.cocodataset.org bucket does not exist

[vincent-assoun]

still down

[pdejorge]

Still down...

[hoeoek]

Still down

[ericskh2]

still down

[ddw2AIGROUP2CQUPT]

still down

[hwaseem04]

I was able to download via curl in the Terminal. For example:

curl -O http://images.cocodataset.org/zips/val2017.zip

This works !! Thanks.

[harakas]

Followed instructions. This is not working:

$ gsutil -m rsync gs://images.cocodataset.org/annotations .
BucketNotFoundException: 404 gs://images.cocodataset.org bucket does not exist.
$ gsutil -m rsync gs://images.cocodataset.org/val2017 val2017
BucketNotFoundException: 404 gs://images.cocodataset.org bucket does not exist.

[nardienapratama]

One workaround is to use wget as described here: https://gist.github.com/mkocabas/a6177fc00315403d31572e17700d7fd9

This works with the trainval2014 annotations - thank you!