Datasets:

gclef-cmu
/

pianobench

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.

PianoBench

Project page · Paper · PianoTalk

PianoBench is a benchmark for evaluating musical perception and reasoning in language models. It contains 250 human-verified multiple-choice questions about short piano passages from instructional videos, covering topics from notes, chords, and rhythm to accompaniment and musical style.

Each passage is provided as a MIDI transcription, synthesized audio rendered with Pianoteq, and a text encoding of the MIDI notes. PianoBench is constructed from the held-out test split of PianoTalk, our dataset of interleaved spoken commentary and piano transcripts.

Download PianoBench

NOTE: Five questions have known transcription or reference-answer errors, marked by known_error: true. The paper's results include all 250 questions. See Appendix C.3 for details.

Benchmark layout

pianobench/
├── manifest.json
└── questions/question_<id>/
    ├── pianoteq.wav
    ├── piano.mid
    ├── piano.document.txt
    └── midi-question.prompt.txt

The benchmark contains 250 questions from 162 source videos.

Asset	Description
`pianoteq.wav`	MIDI rendered with Pianoteq, using the NY Steinway D Classical Recording preset
`piano.mid`	Automatic MIDI transcription for the passage
`piano.document.txt`	Text encoding of the MIDI notes (MIDI-text)
`midi-question.prompt.txt`	Supplied MIDI question prompt

Manifest

manifest.json records transcription and rendering metadata and a questions list. Each question contains:

Field	Description
`id`	Stable question identifier
`youtube_id`	Source YouTube video ID
`start_second`, `end_second`	Clip interval in the source recording, in seconds
`question`	Question text
`choices`	Four answer choices, in evaluation order
`answer_index`	Zero-based index of the reference answer in `choices`
`assets`	Asset paths relative to `pianobench/`
`known_error`	Boolean flag identifying a known transcription or reference-answer error

Citation and license

PianoBench is distributed under CC BY-NC-SA 4.0. If you use PianoBench, please cite the accompanying paper:

@article{bradshaw2026pianobenchevaluatingmusicalperception,
      title={PianoBench: Evaluating Musical Perception and Reasoning in Language Models}, 
      author={Louis Bradshaw and Irmak Bukey and Daniel Chenyu Lin and John Thickstun and Chris Donahue},
      year={2026},
      eprint={2611.03839},
      archivePrefix={arXiv},
      primaryClass={cs.SD},
      url={https://arxiv.org/abs/2611.03839}, 
}

Downloads last month: 4

Total file size:

670 MB

Paper for gclef-cmu/pianobench

PianoBench: Evaluating Musical Perception and Reasoning in Language Models

Paper • 2611.03839 • Published 3 days ago