Update am parquet shard 2026-08-18T21:15:18.874810

Update README.md 2026-08-18T21:15:05.820159

Update metadata.csv 2026-08-18T21:15:03.146215

Update om parquet shard 2026-08-18T06:29:27.065308

Update README.md 2026-08-18T06:29:21.869728

Update metadata.csv 2026-08-18T06:29:19.962171

Update ti parquet shard 2026-08-18T06:28:26.842709

Update README.md 2026-08-18T06:28:22.879853

Update metadata.csv 2026-08-18T06:28:20.255941

Update am parquet shard 2026-08-18T06:21:25.383220

Update README.md 2026-08-18T06:21:09.174029

Update metadata.csv 2026-08-18T06:21:07.686113

Update am parquet shard 2026-08-17T23:35:29.506731

Export 2026-08-17T23:04:58.280909

Point loader at per-language parquet shards (data/train-*.parquet)

Delete data/train-00000-of-00001.parquet with huggingface_hub

Migrate to per-language parquet shards (ti)

Migrate to per-language parquet shards (om)

Migrate to per-language parquet shards (am)

Update native parquet 2026-08-17T22:34:12.697626

Export 2026-08-17T21:34:06.214340

Rewrite parquet with smaller row groups (fix datasets-server 300MB scan limit)

Point dataset loader at native parquet (data/train-00000-of-00001.parquet)

Add native parquet with embedded audio bytes (fix broken audiofolder conversion)

Force reprocess: fix stale statistics worker cache

Export 2026-08-17T08:12:39.652454

Export 2026-08-17T08:09:58.640175

Export 2026-08-16T14:16:41.502437

Export 2026-08-15T22:58:22.422466

Export 2026-08-15T22:04:44.947772

Delete submission_form_data.txt with huggingface_hub

Export 2026-08-15T21:31:03.488175

Export 2026-08-15T17:31:06.770240

Export 2026-08-15T17:28:33.671432

Export 2026-08-14T21:37:09.786873

Export 2026-08-14T18:04:04.740256

Export 2026-08-14T18:03:00.340853

Export 2026-08-14T17:18:40.682210

Export 2026-08-14T17:17:03.835373

Export 2026-08-14T14:30:23.829150

Export 2026-08-14T11:41:59.283898

Export 2026-08-13T09:54:00.142957

Export 2026-08-13T08:05:01.662682

initial commit