search

allenai / satlas-super-resolution

Apache License 2.0
220 stars 24 forks source link

Extracting the train_urban_set #9

Closed SarmadNR closed 9 months ago

SarmadNR commented 9 months ago

Hi, this might be obvious but I am struggling to extract these files. The train_urban_set contains three file

  1. train_urban_set.z01
  2. train_urban_set.z02
  3. train_urban_set.z03

How to extract these? I have tried unzip / tar.

Thanks.

piperwolters commented 9 months ago

What error are you getting when running unzip?

SarmadNR commented 9 months ago

Hi , Thank you for your reply.

https://linuxconfig.org/how-to-split-zip-archive-into-multiple-blocks-of-a-specific-size

I am following the above guide and according to it I need a .zip file that does not exist. Where can I find this .zip file?

On my linux machine I ran the following command:

unzip train_urban_set.z01 train_urban_set.z02 train_urban_set.z03

I get the following error.

Archive: train_urban_set.z01 End-of-central-directory signature not found. Either this file is not a zipfile, or it constitutes one disk of a multi-part archive. In the latter case the central directory and zipfile comment will be found on the last disk(s) of this archive. unzip: cannot find zipfile directory in one of train_urban_set.z01 or train_urban_set.z01.zip, and cannot find train_urban_set.z01.ZIP, period.

piperwolters commented 9 months ago

Ah okay, I'll re-upload these with the .zip file too. In the mean time, if you are able to download this, it contains the train/val/test data all in one and should not have this issue.

SarmadNR commented 9 months ago

Thanks for this link. However, It seems like the same link that is already provided in the repo. It contains 128 x 128 NAIP images. Whereas I am interested in 512 x 512 present in it. I will download it anyway to double-check.

Looking forward to the .zip file.

piperwolters commented 9 months ago

Oh, I'm sorry but we only provide the downsampled NAIP images of shape 128x128. We trained most models to upsample by a factor of 4, since semantic accuracy is already a challenge with 4x upsampling.

SarmadNR commented 9 months ago

Okay, thanks for the information!