File size: 2,858 Bytes
5a7ebd6 cd36994 5a7ebd6 609d268 5a7ebd6 cd36994 5a7ebd6 cd36994 5a7ebd6 245c6f0 5a7ebd6 cd36994 5a7ebd6 cd36994 245c6f0 5a7ebd6 cd36994 5a7ebd6 cd36994 5a7ebd6 cd36994 5a7ebd6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | ---
language:
- sw
license: cc-by-nc-4.0
base_model: facebook/mms-tts-swh
tags:
- text-to-speech
- tts
- vits
- swahili
- mms
datasets:
- google/WaxalNLP
pipeline_tag: text-to-speech
---
# TTSModel — Swahili TTS
A Swahili text-to-speech model, finetuned from Meta's [MMS-TTS Swahili](https://huggingface.co/facebook/mms-tts-swh) checkpoint on the `swa_tts` split of [google/WaxalNLP](https://huggingface.co/datasets/google/WaxalNLP), using the [VITS finetuning recipe](https://github.com/ylacombe/finetune-hf-vits) from `ylacombe/finetune-hf-vits`. Developed by the Maseno Centre for Applied AI (MCAAI).
## Model details
- **Base model:** `facebook/mms-tts-swh` (Meta's Massively Multilingual Speech TTS, Swahili)
- **Architecture:** VITS (single-speaker)
- **Training data:** [google/WaxalNLP](https://huggingface.co/datasets/google/WaxalNLP), `swa_tts` config — 1,387 train utterances (805 after duration/length filtering), single Swahili speaker, 16kHz audio, sourced via the Loud & Clear initiative.
- **Training checkpoint:** step 4,500 / 20,200 planned steps
- **Sample rate:** 16,000 Hz
- **Language:** Swahili (`swh` / ISO 639-3)
## Training configuration
| Setting | Value |
|---|---|
| Learning rate | 2e-5 |
| Batch size | 8 |
| Precision | fp16 |
| Max clip duration | 20s |
| Min clip duration | 0.5s |
| Loss weights | mel=35, kl=1.5, disc=3, gen/fmaps/duration=1 |
Training used the `finetune-hf-vits` recipe with `transformers==4.35.1`, `datasets==2.14.7`, `accelerate==0.24.1`, `numpy<2.0`.
## Usage
```python
import numpy as np
from transformers import pipeline
from IPython.display import Audio as IPyAudio
synthesiser = pipeline("text-to-speech", model="MCAA1-MSU/TTSModel")
speech = synthesiser("Habari yako, karibu Kenya.")
audio = np.squeeze(speech["audio"])
display(IPyAudio(audio, rate=speech["sampling_rate"]))
```
Verified loading cleanly with `transformers` `pipeline("text-to-speech", ...)` — no missing/unexpected weight warnings on load.
## Intended use
Research and experimentation with Swahili TTS.
## Limitations
- **Single speaker, single dataset**: may not generalize well to varied Swahili dialects, accents, or speaking styles.
- **No formal evaluation yet**: no MOS, WER, or MCD scores computed. Qualitative spot-checks suggest intelligible but rough output.
- **Small dataset**: ~800 training utterances after filtering.
## License
Derived from `facebook/mms-tts-swh` (**CC-BY-NC-4.0**, non-commercial). This finetuned model inherits that license. `swa_tts` training data is CC-BY-SA-4.0.
## Acknowledgements
- [Meta MMS](https://huggingface.co/facebook/mms-tts-swh) for the base checkpoint
- [WAXAL / google/WaxalNLP](https://huggingface.co/datasets/google/WaxalNLP) for the training data
- [ylacombe/finetune-hf-vits](https://github.com/ylacombe/finetune-hf-vits) for the finetuning recipe
|