chime-libricount / release.json
ltuncay's picture
Publish source-preserving libricount prepared candidate (part 2)
1223d50 verified
Raw History Blame Contribute Delete
4.75 kB
{
"dataset_id": "libricount",
"files": [
{
"bytes": 11170888,
"path": "manifest.jsonl",
"sha256": "0ffa8237f2d6f68a298bf9dec828389701bb30267448cc37970a4fd7611b074a"
},
{
"bytes": 933580800,
"path": "data/test/shard-000000.tar",
"samples": 5720,
"sha256": "d5c8405309e2023321d115bfb6a08c08d71551b92feb4a7bd39e409de464f179",
"split": "test"
}
],
"format": "webdataset",
"manifest": {
"bytes": 11170888,
"path": "manifest.jsonl",
"sha256": "0ffa8237f2d6f68a298bf9dec828389701bb30267448cc37970a4fd7611b074a"
},
"metadata": {
"benchmark_revision": {
"audio_unchanged": true,
"kind": "chime-v1-usage",
"previous_publication_sha256": "f2165cdbb3bd346028d69efd624cd781762e8ed0dc9367e24d4fecdb69cef13b",
"previous_revision": "7e568b331c2ef0600d131e3223418153e08a5852",
"source_selection_unchanged": true
},
"catalogue_version": "public-acquisition-2026-10-28-v1",
"preparation": {
"adapter_version": 1,
"groups": 5720,
"metadata_joined_examples": 5720,
"native_audio": "unchanged",
"quality_policy": "Exclude invalid targets/payloads only; retain noise, difficulty, overlap, and valid score distributions.",
"splits": {
"test": 5720
}
},
"scientific_admission": "pending",
"source": {
"access_conditions": "Public HTTPS download without registration or signed agreement. Preserve source attribution and license restrictions.",
"citation": "St\u00f6ter, Chakrabarty, Habets & Edler (2018), LibriCount dataset.",
"license": "CC-BY-4.0",
"metadata_url": "https://zenodo.org/api/records/1216072",
"notes": "",
"quality_exclusions": {
"exclude": [
"missing or malformed required target with no recoverable authoritative annotation",
"audio/annotation identifier mismatch or invalid time bounds",
"author-withdrawn examples",
"undecodable or truncated files"
],
"retain": [
"background noise",
"difficulty, accents, child speech or dysarthria severity",
"low agreement when valid annotator distributions remain available",
"overlap within recordings or across datasets; annotate overlap and group related examples"
],
"task_specific": "Mask only invalid task targets when other targets remain valid; do not infer poor-label status from acoustic difficulty."
},
"release": "v1.0.0",
"selection": "Full LibriCount10-0dB release, preserving all speaker-count classes 0\u201310 and source mixture metadata.",
"source_url": "https://zenodo.org/records/1216072",
"subset_policy": {
"deterministic_order": "Preserve official splits; group by speaker/session where supplied, stratify by declared labels/language, order groups by SHA256(\"chime-public-v1\\0\" + group_id), add complete groups round-robin while both byte budgets permit. Record every included/excluded ID and reason.",
"downmix": false,
"expanded_parent_bytes": 917964087,
"extraction": "Stream source archives; materialize only selected native files and required metadata. Do not assemble split source archives or extract the full parent without a separate measured workspace allowance.",
"inventory_evidence": "docs/curation/speech-selections/libricount-archive-bounds.json",
"manifest_required": true,
"native_audio": true,
"pcm_bytes_limit": 917964087,
"resample": false,
"selection": "Full LibriCount10-0dB release, preserving all speaker-count classes 0\u201310 and source mixture metadata.",
"stored_bytes_limit": 917964087
}
},
"source_objects": [
{
"bytes": 832499962,
"checksum": {
"algorithm": "md5",
"value": "30c8f844dc59fa65d216d53db9dc37e2"
},
"dataset": "libricount",
"name": "libricount/LibriCount10-0dB.zip",
"url": "https://zenodo.org/records/1216072/files/LibriCount10-0dB.zip?download=1"
}
]
},
"native_audio_bytes": 915451680,
"pcm_bytes": 915200000,
"sample_count": 5720,
"schema_version": 1,
"scientific_admission": "pending",
"shard_size_bytes": 1000000000,
"shards": [
{
"bytes": 933580800,
"path": "data/test/shard-000000.tar",
"samples": 5720,
"sha256": "d5c8405309e2023321d115bfb6a08c08d71551b92feb4a7bd39e409de464f179",
"split": "test"
}
],
"stage": "prepared-candidate",
"status": "complete",
"validation": {
"audio_sha256": "all_assets",
"shard_readback": "all_members_and_native_headers",
"source_decode": "all_frames_finite"
}
}