---
license: cc-by-4.0
language:
- fr
task_categories:
- automatic-speech-recognition
- text-to-speech
tags:
- lfm2
- lfm2-audio
- unsloth
- mimi
pretty_name: SIWIS French for LFM2.5-Audio (Unsloth tutorial)
---

# SIWIS French for LFM2.5-Audio — Unsloth tutorial data

2000 training clips and 100 validation clips of [SIWIS](https://datashare.ed.ac.uk/handle/10283/2353)
(one professional French female voice, studio recordings; via `Rcarvalo/siwis-fr-24k`, CC BY 4.0), each as two
training records for [LiquidAI/LFM2.5-Audio-1.5B](https://huggingface.co/LiquidAI/LFM2.5-Audio-1.5B):

| `task` | the user | the assistant | `codes` |
|---|---|---|---|
| `asr` | speaks (`audio_bytes`, 24 kHz FLAC, resampled to 16 kHz at training time) | writes `text` | — |
| `tts` | writes `text` | speaks | 8 Mimi codebooks (kyutai/mimi, 24 kHz, 12.5 frames/s) of the clip, plus the end-of-audio frame `2048` |

The split keeps a SIWIS sentence group (`source_file`: part 3 reads each sentence twice) on one side only.
Built by `scripts/build_dataset.py` of the tutorial (seed 3407); 2.20 h of speech in training.

Please cite SIWIS: Honnet, P.-E., Lazaridis, A., Garner, P. N., Yamagishi, J. (2017). The SIWIS French Speech
Synthesis Database. Idiap-RR-02-2017. License of the original data: CC BY 4.0.