huggingface / huggingface/datasets
Image Data loaded Twice
- Dominant language
- Python
- Stars
- 22k
- Forks
- 3.4k
- Avg merge
- 5d 7h
- Merged PRs (30d)
- 17
Description
### Describe the bug

When I learn from https://huggingface.co/docs/datasets/image_load and try to load image data from a folder. I noticed that the image was read twice in the returned data. As you can see in the attached image, there are only four images in the train folder, but reading brings up eight images
### Steps to reproduce the bug
from datasets import Dataset, load_dataset
dataset = load_dataset("imagefolder", data_dir="data/", drop_labels=False)
# print(dataset["train"][0]["image"] == dataset["train"][1]["image"])
print(dataset)
print(dataset["train"]["image"])
print(len(dataset["train"]["image"]))
### Expected behavior
DatasetDict({
train: Dataset({
features: ['image', 'label'],
num_rows: 8
})
})
[, , , , , , , ]
8
### Environment info
- `datasets` version: 2.14.5
- Platform: Windows-10-10.0.22621-SP0
- Python version: 3.9.17
- Huggingface_hub version: 0.19.4
- PyArrow version: 13.0.0
- Pandas version: 2.0.3
Contributor guide
Assessment
This issue has not been assessed yet.