Datasets:

---

HuggingFaceFW

/

finepdfs

like 636

Follow

FineData 1.03k

I tried to load the main English subset of the dataset as:

dataset = load_dataset("HuggingFaceFW/finepdfs", name="eng_Latn",
                           split="train", num_proc=8)

It downloads successfully, but fails during data "generation". I tried it across two different machines, and also without the num_proc argument.
Error traceback:

Generating train split: 0 examples [00:02, ? examples/s]████████████████████████████████████████████████████████████████▎                                                      | 32/50 [00:00<00:00, 119.09files/s]
multiprocess.pool.RemoteTraceback:
"""
Traceback (most recent call last):
  File "/home/anshumansuri/miniconda3/envs/pd/lib/python3.12/site-packages/datasets/builder.py", line 1854, in _prepare_split_single
    for _, table in generator:
                    ^^^^^^^^^
  File "/home/anshumansuri/miniconda3/envs/pd/lib/python3.12/site-packages/datasets/packaged_modules/parquet/parquet.py", line 93, in _generate_tables
    for batch_idx, record_batch in enumerate(
                                   ^^^^^^^^^^
  File "pyarrow/_dataset.pyx", line 3830, in _iterator
  File "pyarrow/_dataset.pyx", line 3436, in pyarrow._dataset.TaggedRecordBatchIterator.__next__
  File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
  File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
OSError: Repetition level histogram size mismatch

I am on datasets==3.3.2