#285
/capstor/store/cscs/swissai/infra01/audio-datasets/tokenized/sft/marco_longspeech_sft
| kind | tokenized |
|---|---|
| status | active |
| samples | 204,881 |
| counted via | megatron idx |
| size | 22.0 GB |
| files | 18 |
| first seen | 2026-07-22 13:36 |
| last seen | 2026-07-22 13:36 |
| registered | 2026-07-22 13:36 |
# dataset #285 path: /capstor/store/cscs/swissai/infra01/audio-datasets/tokenized/sft/marco_longspeech_sft kind: tokenized | status: active samples: 204,881 (megatron idx) | size: 22.0 GB | files: 18 tree (truncated): ``` _STAGE_MANIFEST.json _SUCCESS lct/rank_0000_chunk_0000.bin lct/rank_0000_chunk_0000.cut_ids.jsonl.zst lct/rank_0000_chunk_0000.idx lct/rank_0001_chunk_0000.bin lct/rank_0001_chunk_0000.cut_ids.jsonl.zst lct/rank_0001_chunk_0000.idx lct/rank_0002_chunk_0000.bin lct/rank_0002_chunk_0000.cut_ids.jsonl.zst lct/rank_0002_chunk_0000.idx lct/rank_0003_chunk_0000.bin lct/rank_0003_chunk_0000.cut_ids.jsonl.zst lct/rank_0003_chunk_0000.idx lct/rank_0004_chunk_0000.bin lct/rank_0004_chunk_0000.cut_ids.jsonl.zst lct/rank_0004_chunk_0000.idx sft_materialize_summary.json ```