RecursionError using tfds.load to import tensorflow-dataset (Mac)

[valeriogiampa]

Short description tfds.load function does not work properly.

Environment information

• Operating System: Mac OS Sonoma 14.5 (tested also on Windows 11)

• Python version: 3.12.3

• tensorflow-datasets version: 4.9.4

• tensorflow version: 2.16.1

• Does the issue still exists with the last tfds-nightly package (pip install --upgrade tfds-nightly) ? I don't know. This is the first version installed on my pc.

Reproduction instructions

import tensorflow as tf
import sklearn
import tensorflow_datasets as tfds

print(tf.__version__)
print(sklearn.__version__)

# Load the Iris dataset
(data), info = tfds.load('iris', with_info=True, split='train')

Link to logs the log is in the attached file log.docx

Expected behavior What you expected to happen.

Additional context The code is the same available in google documentation and inside the Google Cloud Machine Learning didactic materials from EICTA.

[pierrot0]

Hi Valerio,

Would you mind sharing the stack-trace and error you are getting please? The code you provided doesn't lead to any errors on my side.

I tested with both TFDS 4.9.4 and nightly.

Thanks, Pierre

[valeriogiampa]

I have uploaded the screenshot of the executed code using python and using Jupyter.

Log2.docx