datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
news-segmentationspeech_commands_pitchmlcs_en_pitchPitchBench
PitchBench
A benchmark for testing what audio / acoustic signals Audio Language Models (ALMs) do
and don't understand. PitchBench probes pitch perception across 29 controlled
experiments — single-pitch ID, onsets/offsets, chords, sequences, contour, audio
effects, and polyphonic streams.
Each row is one (audio, question, answer) triple: a short WAV stimulus, the
question (prompt*) asked of the model, and the ground-truth answer fields
(experiment-specific column names).… See the full description on the dataset page: https://huggingface.co/datasets/pitchbench-authors/PitchBench.NSYNTH_PITCH_HEARmlcs_fr_pitchbenchmark-pitchspeech_commands_pitch_300hzv14-fake-fast-pitchmlcs_et_pitch43-143-phase2-appconv-ime-high_pitchspeech_commands_pitch_200hzmlcs_rw_pitchinterleaving_mmau_lowest_pitchspeech_mmau_highest_pitchmlcs_de_pitchpitch_ratepitch-benchmark
Pitch Benchmark
Evaluation and training data for the pitch tracker benchmark at
https://github.com/lars76/pitch-benchmark. Two archives:
File
Size
Contents
eval.tar
382 MB
ten evaluation corpora, test and calibration clips, nine renderings each, reference labels, manifests
train.tar
4.5 GB
five training corpora with labels and the augmentation pools
Checksums are in SHA256SUMS.
Use
hf download lars1234/pitch-benchmark --repo-type dataset… See the full description on the dataset page: https://huggingface.co/datasets/lars1234/pitch-benchmark.43-143-phase2-appconv-ime-low_pitchmlcs_en_pitch_uniform_6Pitch_Level_Distinctionspeech_commands_pitch_250hzPitchExtractionByLyrics_CSDcommon_voice_16_0_ko_pitchspeech_commands_pitch_100hzspeech_commands_pitch_350hzspeech_commands_pitch_150hzspeech_mmau_lowest_pitchspeech_commands_pitch_50hzmlcs_en_pitch_uniform_4