Reputation: 384
I want to build a model in tensorflow 2.0 using the 'celeb_a' dataset in Spyder (Anaconda3).
When running this code:
celeba_bldr = tfds.builder('celeb_a')
celeba_bldr.download_and_prepare()
I get the following error:
NonMatchingChecksumError: Artifact https://drive.google.com/uc?export=download&id=0B7EVK8r0v71pZjFTYXZWM3FlRnM, downloaded to C:\Users\Home\tensorflow_datasets\downloads\ucexport_download_id_0B7EVK8r0v71pZjFTYXZWM3FlDDaXUAQO8EGH_a7VqGNLRtW52mva1LzDrb-V723OQN8.tmp.3706deb2f0c24dcfbc0d2de880c8aa70\uc, has wrong checksum.
I have no idea how to overcome this error.
Does anybody have a solution for this error so that I can download the 'celeb_a' dataset?
Upvotes: 6
Views: 4634
Reputation: 4100
Try updating tensorflow-datasets
.
Anaconda is very outdated having version 1.2 while PIP already has 4.4.
Upvotes: 1
Reputation: 100
Sometimes server was bad due to which it fails to take requests from data server that's why you are getting this error, I tried it now it works fine see this colab notebook.
Also I tried to register checksum so that if there is any problem with that but it is correct see colab . So problem occur due to bad server. Try it again it works.
Upvotes: 1