avey-b-ur / MANIFEST.json
mahwizzzz's picture
Publish Avey-B Urdu 25M with benchmarks and model card
4cae2c2 verified
Raw History Blame Contribute Delete
1.35 kB
{
"generated_at": "2026-09-10T09:59:10.314711+00:00",
"lang": "urd_Arab",
"map_url": "https://data.hplt-project.org/three/sorted/urd_Arab.map",
"bins_requested": [
10,
9
],
"prob_threshold": 0.8,
"min_chars": 20,
"val_fraction": 0.005,
"max_docs_per_shard": 50000,
"max_chars_per_shard": 500000000,
"max_shards": null,
"stream_download": true,
"shards": [
{
"url": "https://data.hplt-project.org/three/sorted/urd_Arab/10_1.jsonl.zst",
"bin": 10,
"raw_docs": 2162,
"dropped_low_prob": 0,
"dropped_empty": 0,
"dropped_exact_dup": 0,
"pii_spans_redacted": 24,
"kept_train": 2151,
"kept_val": 11,
"chars_kept": 41055559
},
{
"url": "https://data.hplt-project.org/three/sorted/urd_Arab/9_1.jsonl.zst",
"bin": 9,
"raw_docs": 29290,
"dropped_low_prob": 3,
"dropped_empty": 0,
"dropped_exact_dup": 1,
"pii_spans_redacted": 729,
"kept_train": 29146,
"kept_val": 140,
"chars_kept": 500109810
}
],
"totals": {
"raw_docs": 31452,
"kept_train": 31297,
"kept_val": 151,
"dropped_low_prob": 3,
"dropped_empty": 0,
"dropped_exact_dup": 1,
"pii_spans_redacted": 753,
"chars_kept": 541165369
},
"consolidated_counts": {
"train": 31297,
"val": 151
}
}