vaughankraska/

imagenet_sketch

This dataset was duplicated from songweig/imagenet_sketch so I could create validation splits and use parquets for caching the dataset instead of requiring python code execution. The dataset was replicated using the following script: from datasets import load_dataset, DatasetDict import os

def main(): ds = load_dataset("songweig/imagenet_sketch", split="train").shuffle(seed=42)

ds_train = ds.select(range(40_000))
ds_validate =… See the full description on the dataset page: https://huggingface.co/datasets/vaughankraska/imagenet_sketch.