animalspeak-pseudovox / croissant.json
chrispla's picture
Upload Croissant metadata
d277439 verified
{
"@context": {
"@vocab": "https://schema.org/",
"sc": "https://schema.org/",
"cr": "http://mlcommons.org/croissant/",
"rai": "http://mlcommons.org/croissant/RAI/",
"prov": "http://www.w3.org/ns/prov#"
},
"@type": "sc:Dataset",
"name": "AnimalSpeak Pseudovox Train-Unseen",
"description": "Train-unseen split of AnimalSpeak Pseudovox: short, silence-trimmed, single-vocalization clips with per-clip metadata.",
"url": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox",
"version": "1.0.0",
"datePublished": "2026-05-06",
"license": "CC-BY-NC-4.0",
"isAccessibleForFree": false,
"isLiveDataset": false,
"keywords": [
"bioacoustics",
"animal vocalizations",
"single vocalization",
"AnimalSpeak",
"Pseudovox",
"train unseen"
],
"creator": [
{
"@type": "Organization",
"name": "Earth Species Project"
}
],
"publisher": {
"@type": "Organization",
"name": "Earth Species Project",
"url": "https://www.earthspecies.org/"
},
"distribution": [
{
"@type": "cr:FileObject",
"@id": "metadata-parquet",
"name": "Per-example metadata",
"description": "Parquet table with one row per audio clip (346,907 rows).",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/metadata.parquet",
"encodingFormat": "application/vnd.apache.parquet",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000000",
"name": "WebDataset shard 000000",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000000.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000001",
"name": "WebDataset shard 000001",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000001.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000002",
"name": "WebDataset shard 000002",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000002.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000003",
"name": "WebDataset shard 000003",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000003.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000004",
"name": "WebDataset shard 000004",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000004.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000005",
"name": "WebDataset shard 000005",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000005.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000006",
"name": "WebDataset shard 000006",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000006.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000007",
"name": "WebDataset shard 000007",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000007.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000008",
"name": "WebDataset shard 000008",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000008.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000009",
"name": "WebDataset shard 000009",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000009.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000010",
"name": "WebDataset shard 000010",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000010.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000011",
"name": "WebDataset shard 000011",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000011.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000012",
"name": "WebDataset shard 000012",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000012.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000013",
"name": "WebDataset shard 000013",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000013.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000014",
"name": "WebDataset shard 000014",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000014.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000015",
"name": "WebDataset shard 000015",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000015.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000016",
"name": "WebDataset shard 000016",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000016.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileObject",
"@id": "shard-000017",
"name": "WebDataset shard 000017",
"description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
"contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000017.tar",
"encodingFormat": "application/x-tar",
"prov:wasGeneratedBy": {
"@id": "activity-hf-export"
}
},
{
"@type": "cr:FileSet",
"@id": "audio-file-set",
"name": "WAV audio files inside WebDataset shards",
"description": "16 kHz WAV clips stored inside tar shards under audio/.",
"containedIn": [
{
"@id": "shard-000000"
},
{
"@id": "shard-000001"
},
{
"@id": "shard-000002"
},
{
"@id": "shard-000003"
},
{
"@id": "shard-000004"
},
{
"@id": "shard-000005"
},
{
"@id": "shard-000006"
},
{
"@id": "shard-000007"
},
{
"@id": "shard-000008"
},
{
"@id": "shard-000009"
},
{
"@id": "shard-000010"
},
{
"@id": "shard-000011"
},
{
"@id": "shard-000012"
},
{
"@id": "shard-000013"
},
{
"@id": "shard-000014"
},
{
"@id": "shard-000015"
},
{
"@id": "shard-000016"
},
{
"@id": "shard-000017"
}
],
"includes": "audio/*.wav",
"encodingFormat": "audio/wav",
"prov:wasGeneratedBy": [
{
"@id": "activity-source-separation"
},
{
"@id": "activity-window-extraction"
},
{
"@id": "activity-hf-export"
}
]
}
],
"recordSet": [
{
"@type": "cr:RecordSet",
"@id": "metadata",
"name": "Metadata rows",
"description": "One row per Pseudovox clip. The audio_name column refers to the WAV entry stored under audio/ inside a tar shard.",
"key": [
{
"@id": "metadata/id"
}
],
"field": [
{
"@type": "cr:Field",
"@id": "metadata/id",
"name": "id",
"description": "Stable clip identifier, derived from the WAV filename stem.",
"dataType": "sc:Text",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "id"
}
}
},
{
"@type": "cr:Field",
"@id": "metadata/audio_name",
"name": "audio_name",
"description": "WAV filename inside shard audio/ directories.",
"dataType": "sc:Text",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "audio_name"
}
}
},
{
"@type": "cr:Field",
"@id": "metadata/duration_s",
"name": "duration_s",
"description": "Clip duration in seconds.",
"dataType": "sc:Float",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "duration_s"
}
},
"prov:wasGeneratedBy": {
"@id": "activity-window-extraction"
}
},
{
"@type": "cr:Field",
"@id": "metadata/species_scientific_gbif",
"name": "species_scientific_gbif",
"description": "Scientific species name when available.",
"dataType": "sc:Text",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "species_scientific_gbif"
}
},
"prov:wasGeneratedBy": {
"@id": "activity-species-labeling"
}
},
{
"@type": "cr:Field",
"@id": "metadata/custom_classifier_species",
"name": "custom_classifier_species",
"description": "Top species predicted by the custom classifier when available.",
"dataType": "sc:Text",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "custom_classifier_species"
}
},
"prov:wasGeneratedBy": {
"@id": "activity-species-labeling"
}
},
{
"@type": "cr:Field",
"@id": "metadata/custom_classifier_max_p",
"name": "custom_classifier_max_p",
"description": "Custom classifier maximum probability for the predicted species.",
"dataType": "sc:Float",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "custom_classifier_max_p"
}
},
"prov:wasGeneratedBy": {
"@id": "activity-species-labeling"
}
},
{
"@type": "cr:Field",
"@id": "metadata/birdnet_confidence",
"name": "birdnet_confidence",
"description": "BirdNET confidence score.",
"dataType": "sc:Float",
"source": {
"fileObject": {
"@id": "metadata-parquet"
},
"extract": {
"column": "birdnet_confidence"
}
},
"prov:wasGeneratedBy": {
"@id": "activity-birdnet-scoring"
}
}
],
"examples": [
{
"id": "Amazona_rhodocorytha_SSS_XC3_Jandaira_source1_clip1",
"audio_name": "Amazona_rhodocorytha_SSS_XC3_Jandaira_source1_clip1.wav",
"duration_s": 0.37615625
}
]
},
{
"@type": "cr:RecordSet",
"@id": "audio-files",
"name": "Audio files",
"description": "Audio objects extracted from the tar shards.",
"field": [
{
"@type": "cr:Field",
"@id": "audio-files/path",
"name": "path",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "audio-file-set"
},
"extract": {
"fileProperty": "fullpath"
}
}
},
{
"@type": "cr:Field",
"@id": "audio-files/content",
"name": "content",
"dataType": "sc:AudioObject",
"source": {
"fileSet": {
"@id": "audio-file-set"
},
"extract": {
"fileProperty": "content"
}
}
}
]
}
],
"measurementTechnique": [
"BirdMixIT source separation",
"DRASDIC-style short activity-window extraction",
"sl-efficientnet-bio species prediction and metadata-match filtering",
"BirdNET confidence scoring"
],
"variableMeasured": [
{
"@type": "PropertyValue",
"name": "duration_s",
"description": "Clip duration in seconds."
},
{
"@type": "PropertyValue",
"name": "species_scientific_gbif",
"description": "Scientific species name when available."
},
{
"@type": "PropertyValue",
"name": "custom_classifier_max_p",
"description": "Confidence of the custom species classifier."
},
{
"@type": "PropertyValue",
"name": "birdnet_confidence",
"description": "BirdNET confidence score."
}
],
"usageInfo": {
"downstreamUse": "Research on bioacoustic representation learning, vocalization-level audio-language training data, and evaluation of held-out species generalization.",
"outOfScopeUse": "Do not use as the sole basis for ecological monitoring, sensitive species localization, or unsupported claims about real-world species presence.",
"biasRisksLimitations": {
"bias": "Coverage reflects the source recording collection, train-unseen split construction, and classifier confidence filters.",
"risks": "Models trained on this data can inherit taxonomic, geographic, and recording-quality biases.",
"limitations": "Clips are short isolated vocalizations and may not represent full soundscape context, overlapping callers, or field deployment conditions."
},
"safetyMitigations": "Validate on real target-domain recordings, document target taxa and recording conditions, and consult bioacoustics expertise for high-stakes uses."
},
"rai:dataLimitations": "Short isolated Pseudovox clips do not capture all ecological and acoustic context present in the original recordings.",
"rai:dataBiases": "Taxonomic and acoustic coverage is inherited from source recordings, split criteria, and classifier confidence thresholds.",
"rai:personalSensitiveInformation": "No human personal data is intentionally included. Source recordings may contain incidental metadata or background sounds inherited from the source collection.",
"rai:dataUseCases": "Bioacoustic model research, training data construction, and held-out species experiments.",
"rai:dataSocialImpact": "The dataset may support better animal-sound AI systems, but ecological deployment requires expert validation and safeguards for sensitive species and sites.",
"rai:hasSyntheticData": false,
"prov:wasDerivedFrom": [
{
"@type": "sc:Dataset",
"name": "AnimalSpeak"
}
],
"prov:wasGeneratedBy": [
{
"@type": "prov:Activity",
"@id": "activity-source-separation",
"name": "Source separation",
"description": "Original field recordings were source-separated with BirdMixIT to isolate candidate animal vocalization sources.",
"prov:used": [
{
"@type": "prov:SoftwareAgent",
"name": "BirdMixIT"
}
]
},
{
"@type": "prov:Activity",
"@id": "activity-window-extraction",
"name": "Short activity-window extraction",
"description": "Short vocal activity windows were extracted from the separated sources following the DRASDIC pseudovox procedure.",
"prov:wasInformedBy": {
"@type": "sc:CreativeWork",
"name": "DRASDIC"
}
},
{
"@type": "prov:Activity",
"@id": "activity-species-labeling",
"name": "Species prediction and metadata-match filtering",
"description": "Each candidate clip was tagged with a species label predicted by sl-efficientnet-bio, then filtered to retain clips whose predicted species matched the recording metadata.",
"prov:used": [
{
"@type": "prov:SoftwareAgent",
"name": "sl-efficientnet-bio"
}
]
},
{
"@type": "prov:Activity",
"@id": "activity-birdnet-scoring",
"name": "BirdNET confidence scoring",
"description": "BirdNET confidence scores were computed for released clips.",
"prov:used": [
{
"@type": "prov:SoftwareAgent",
"name": "BirdNET"
}
]
},
{
"@type": "prov:Activity",
"@id": "activity-hf-export",
"name": "Hugging Face sharded export",
"description": "The filtered train-unseen split was packaged as WebDataset tar shards, metadata.parquet, and Croissant JSON-LD for this Hugging Face release."
}
],
"conformsTo": "http://mlcommons.org/croissant/1.1"
}