Datasets:

yuvalkirstain

/

pickapic_v1

like 32

I've tried to:

• set parameter split to "test_unique" in datasets.load_dataset(), which seemed to lead to downloading the whole dataset anyway.
• turn to https://huggingface.co/datasets/yuvalkirstain/pickapic_v1_no_images , which got 404 (on 2023-5-5)
• download from https://huggingface.co/datasets/yuvalkirstain/pickapic_v1/blob/main/data/test_unique-00000-of-00001-c163d5839d8aed9f.parquet to get the test_unique split, but failed in reading it with pandas dut to an error that I haven't be able to solve so far:

ArrowInvalid: Could not open Parquet input source '': Parquet magic bytes not found in footer. Either the file is corrupted or this is not a parquet file.

• visit /parquet endpoint in Datasets-Server, which returned:

{
    "parquet_files": [],
    "pending": [],
    "failed": [
        {
            "kind": "config-parquet",
            "dataset": "yuvalkirstain/pickapic_v1",
            "config": "yuvalkirstain--pickapic_v1",
            "split": null
        }
    ]
}

Could anybody kindly offer some help about any method mentioned above or other methods to download the dataset without images or part of it like test split only?