Nov 5 Today 01:02 Update 2026-11-05: Add 50 new multilingual interpreter interview videos (518 total rows, 16kHz mono WAV) 00:29 Update 2026-11-05: Add 50 new multilingual interpreter interview videos (518 total rows, 16kHz mono WAV) Nov 4 Yesterday 00:14 Update 2026-11-04 daily batch: +50 multilingual interpreter interview videos with 16kHz mono WAV (total 468 rows) Nov 3 2 days ago 00:08 Update batch 2026-11-03: 50 new bilingual interview videos with 16kHz WAV Nov 2 3 days ago 00:10 Update parquet dataset with 50 new embedded 16kHz WAV audio samples (total 368) 00:10 Update videos metadata JSONL (368 videos) Nov 1 4 days ago 00:11 Update 2026-11-01: Add 50 new interpreter interview videos (total 318) Oct 31 5 days ago 00:05 Revert: Restore data/cabin_interpreter_interviews_test.parquet to aacef79 00:05 Revert: Restore data/videos.jsonl to aacef79 00:03 Update 2026-10-31: Add 50 new interpreter interview videos (total 318) 00:03 Update 2026-10-31: Add 50 new interpreter interview videos (total 318) Oct 30 6 days ago 00:07 Cập nhật Parquet hợp nhất 268 mẫu kiểm thử (2026-10-30) 00:07 Cập nhật videos.jsonl 268 dòng (2026-10-30) Oct 29 7 days ago 08:59 Upload data/videos.jsonl with huggingface_hub 08:58 Upload data/cabin_interpreter_interviews_test.parquet with huggingface_hub 08:44 Cập nhật 118 mẫu âm thanh 16kHz mono ngày 29/10/2026 Oct 28 8 days ago 10:47 Update dataset with 68 pre-downloaded multilingual interpreter interview videos 08:39 Restore 76bfa5a6 (README, curated JSON, parquet were overwritten by 53a6596e by mistake); add verified_turns subset in verified/ 08:30 v2: 27 verified interpreter-interview clips (+22), fixed v1 metadata, reference turns, production results 2026-10-28 08:04 Upload data/cabin_interpreter_interviews_test.parquet with huggingface_hub 08:00 Upload data/cabin_interpreter_interviews_test.parquet with huggingface_hub 07:05 README: restore the 52-video card (overwritten by aa009f4 ) and add the per-turn labels section; videos.jsonl: in_audio_parquet instead of v1 fields 06:59 v2: 12 more videos, per-turn language + transcript in a 3-min window per video, auto turns, v1 language fixes, rehydrate.py (no new audio: standard-licence YouTube) 06:47 Upload data/cabin_interpreter_interviews_test.parquet with huggingface_hub 06:30 Update data/cabin_interpreter_interviews_test.parquet: 52 curated interview videos with pre-downloaded audio 06:10 Upload data/cabin_interpreter_interviews_test.parquet with huggingface_hub 
