Open yepw opened 4 months ago
It seems you cannot disable GCS for this dataset as it downloads all files from the buckets (source). Can you download the dataset from GCS and build it using Beam?
I am running into the same issue with the latest TF nightly but this seems unrelated to TFDS. I opened an issue for this on upstream TF https://github.com/tensorflow/tensorflow/issues/69789
@marcenacp Could you have a look at the upstream issue and forward it to the relevant TF team?
Short description Error when loading dataset "berkeley_autolab_ur5". I didn't run it on Colab but tried to download the dataset locally. I tried to disable GCS following the comments here
Environment information
Operating System: Ubuntu 20.04
Python version: 3.11.8
tensorflow-datasets
/tfds-nightly
version: 4.9.4+nightlytensorflow
/tf-nightly
version: 2.17.0-dev20240409Does the issue still exists with the last
tfds-nightly
package (pip install --upgrade tfds-nightly
) ? YesReproduction instructions
Link to logs
Expected behavior It starts downloading the dataset.