Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Passing numpy array to ClassLabel names causes ValueError

See original GitHub issue

Describe the bug

If a numpy array is passed to the names argument of ClassLabel, creating a dataset with those features causes an error.

Steps to reproduce the bug

https://colab.research.google.com/drive/1cV_es1PWZiEuus17n-2C-w0KEoEZ68IX

TLDR:

If I define my classes as:

my_classes = np.array(['one', 'two', 'three'])

Then this errors:

features = Features({'value': Value('string'), 'label': ClassLabel(names=my_classes)})
dataset = Dataset.from_list(my_data, features=features)

ValueError                                Traceback (most recent call last)
[<ipython-input-8-a8a9d53ec82f>](https://localhost:8080/#) in <module>
----> 1 dataset = Dataset.from_list(my_data, features=features)

11 frames
[/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py](https://localhost:8080/#) in _asdict_inner(obj)
    183             for f in fields(obj):
    184                 value = _asdict_inner(getattr(obj, f.name))
--> 185                 if not f.init or value != f.default or f.metadata.get("include_in_asdict_even_if_is_default", False):
    186                     result[f.name] = value
    187             return result

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

But this works:

features2 = Features({'value': Value('string'), 'label': ClassLabel(names=list(my_classes))})
dataset2 = Dataset.from_list(my_data, features=features2)

Expected behavior

If I provide a numpy array of class names, I would expect either an error that the names list is the wrong type, or for it to be cast internally.

Environment info

datasets version: 2.7.1
Platform: Linux-5.15.0-56-generic-x86_64-with-glibc2.10
Python version: 3.8.15
PyArrow version: 10.0.1
Pandas version: 1.5.2

Additionally:

Numpy version: 1.23.5

Issue Analytics

State:
Created 10 months ago
Comments:5 (4 by maintainers)

Top GitHub Comments

1reaction

mariosaskocommented, Dec 9, 2022

Hi! No, I don’t think so. The names parameter is annotated as List[str] (NumPy arrays are not lists), and considering that type checking is not a common practice in Python, I think we can leave the code as-is.

0reactions

albertvillanovacommented, Dec 14, 2022

What about checking for Sequence instead? I think users can pass a list or a tuple as well.

Top Results From Across the Web

ValueError: Data must be passed in as a list of numpy arrays

I am running the simple code below however running into the error with the value error. Then I tried to fix it with...

Main classes - Hugging Face

name (str) — Column name. column (list or np.array) — Column data to be added. Returns ... You can define a sharded dataset...

sklearn.preprocessing.MultiLabelBinarizer

Set to True if output binary array is desired in CSR sparse format. ... A common mistake is to pass in a list,...

Release 0.9.7 Snorkel Team

Raises ValueError – If a specified standard metric is not found in the METRICS ... Parameters *y – A list of np.ndarray of...

Subclassing ndarray — NumPy v1.24 Manual

Besides the additional complexities of subclassing a NumPy array, ... class C(np.ndarray): pass >>> # create a standard ndarray >>> arr = np.zeros((3,)) ......