| --- |
| 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 |
| --- |
| |
| # waxal_swahili-tts-mms |
| |
| 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`. |
|
|
|
|
| ## 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 (~22 epochs of ~805 examples) |
| - **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` (this specific combination was required to avoid dependency-incompatibility crashes on Colab). |
|
|
| ## Usage |
|
|
| ```python |
| import numpy as np |
| from transformers import pipeline |
| import scipy.io.wavfile |
| |
| synthesiser = pipeline("text-to-speech", model="Mwau/waxal_swahili-tts-mms") |
| speech = synthesiser("Habari yako, karibu Kenya.") |
| |
| audio = np.squeeze(speech["audio"]) # squeeze extra dims before saving/playing |
| scipy.io.wavfile.write("output.wav", rate=speech["sampling_rate"], data=audio) |
| ``` |
|
|
| Verified loading cleanly with `transformers` `pipeline("text-to-speech", ...)` — no missing/unexpected weight warnings on load. |
|
|
| ## Intended use |
|
|
| Research and experimentation with Swahili TTS, particularly cross-lingual/low-resource TTS transfer techniques. Not yet suitable for production or user-facing applications given the early training stage. |
|
|
| ## Limitations |
|
|
| - **Single speaker, single dataset**: trained on one speaker's recordings from one dataset; may not generalize well to varied Swahili dialects, accents, or speaking styles. |
| - **No formal evaluation yet**: no MOS, WER, or MCD scores have been computed for this checkpoint. Qualitative spot-checks against WAXAL's held-out `test` split suggest intelligible but rough output. |
| - **Small dataset**: ~800 training utterances after filtering is a small corpus for TTS; may overfit to the training speaker's specific vocal characteristics. |
|
|
| ## License |
|
|
| This model is derived from `facebook/mms-tts-swh`, which is released under **CC-BY-NC-4.0** (non-commercial). This finetuned model inherits that license. The `swa_tts` training data is released under 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 |
|
|
