download_assets.sh

[loujinming147]

Great works! But when I try to run download_assets.sh to download ImageNet-256x256 and ImageNet-512x512 files, I get the following error and the address I get is:
https://slurm-ord.s3.amazonaws.com/datasets/imagenet_256_latent_lmdb/train/data.mdb
https://slurm-ord.s3.amazonaws.com/datasets/imagenet-wds/latent_imagenet_512_train-0000.tar

<Error>
<Code>AccessDenied</Code>
<Message>Access Denied</Message>
<RequestId>7G2B5AN15R4ZJGXZ</RequestId>
<HostId>YK/P2VRZoUArpdown/QNAy3z8t8UTCzN7Gqmdsrp2ntH7aw4VTm9EVYwnhDvpsE0Wu3yT1Eyytow0iYkZ5Z1mR/34iL5SpqX</HostId>
</Error>

[Andy1621]

Same question

[devzhk]

Hi,

The download links have been updated. The latest links are as follows

1. Imagenet256: https://hzpublic-hub.s3.us-west-2.amazonaws.com/maskdit/data/imagenet_256_latent_lmdb/train/
2. Imagenet512: https://hzpublic-hub.s3.us-west-2.amazonaws.com/maskdit/data/imagenet-wds/

Updating the _url_dict in download_assets.py accordingly will fix the issue .

[Andy1621]

Thanks!

[arghavan-kpm]

Hi, the new links are also giving the same access denied errors. Is there any alternative download link? Thanks.

[devzhk]

Hi,

The new links attached above are part of the download URLs. The actual URL for each file is defined in download_assets.py. You can either

1. Update the _url_dict in download_assets.py with the new links and run download_assets.py to download the data.
2. Update the _url_dict, extract the exact URL for each file from download_assets.py, and paste it into your browser to download it.

[arghavan-kpm]

Got it, thanks! We should actually replace these with links for imagenet256-latent-lmdb and imagenet512-latent-wds, respectively. Also, could share updated link for vae as well?

[devzhk]

Yes, it's shared in #14 .