The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/webdataset/webdataset.py", line 84, in _split_generators
raise ValueError(
...<2 lines>...
)
ValueError: The TAR archives of the dataset should be in WebDataset format, but the files in the archive don't share the same prefix or the same types.
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
HPC documentation open corpus
The openly licensed part of the documentation we collected to build ClusterPilot-4B, an open AI assistant for supercomputer users (Team Alembic, EuroHPC AI Factories hackathon, October 2026). Plain text extracted from public HPC, Slurm, MPI, GPU and European supercomputing documentation, one page per row, with its URL and licence.
| File | Rows | Licences |
|---|---|---|
corpus/open.jsonl.gz |
10,032 pages | CC-BY-4.0, CC0, Apache-2.0, MIT, BSD, MPICH (permissive), US public domain |
corpus/sharealike.jsonl.gz |
798 pages | CC-BY-SA and GPL: reuse only under the same licence |
corpus/SOURCES.csv |
10,830 | url, title, licence and licence basis of every page (attribution) |
eval_results.tar.gz |
1,416 files | our evaluation score and summary files from Leonardo (no test-question text) |
Row fields: url, title, license, license_basis, retrieved_at, source_id, text.
How the pages were chosen. Only pages that were downloaded successfully (not robots-disallowed) and whose source licence allows redistribution. Left out: proprietary documentation (for example Intel, NVIDIA), NonCommercial licences, and every site with no stated licence (53,189 pages). From the EU AI Factory sites only LUMI and CSC documentation is included (CC-BY-4.0).
Licences and attribution. Each page keeps its own licence: when you reuse a page, credit its source (url,
title) and follow its license. The licence tags were recorded automatically when the pages were downloaded
(license_basis says from where) and were not reviewed by hand; check the source page before relying on a tag.
The text is an automatic extraction (HTML/PDF to Markdown) and can contain extraction errors.
The question-and-answer data we generated from these documents is in VaibhavKKM/hpc-assistant-qa.
- Downloads last month
- 28