The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider removing the loading script and relying on automated data support (you can use convert_to_parquet from the datasets library). If this is not possible, please open a discussion for direct help.

Load the Pascal Part dataset from the local directory.

Usage:

from datasets import load_dataset
DATA_DIR = "/path/to/PascalPart"
split = "trainval"  # choose from ["trainval"]
ds = load_dataset("yiqun/PascalPart", data_dir=DATA_DIR, split=split)

Structure of DATA_DIR:

|_ Annotations_Part
  |_ # images
|_ VOCdevkit
  |_ VOC2010
    |_ JPEGImages
    |_ ...
Downloads last month
0