rnjema-unima/mms-tts-pul-baseline

WAXAL MMS-TTS — Pular (pul)

Fine-tuning-ready checkpoint for Pular (pul).

WAXAL dataset config	google/WaxalNLP — pul_tts
Data provider	AIMS Senegal
WAXAL data license	CC-BY-SA-4.0
Base model	facebook/mms-tts-ful
Model license	CC-BY-NC 4.0 (MMS base; governs fine-tuned model)

ISO code note

WAXAL uses pul (macro-language / regional variety code). MMS uses ful (specific variety, ISO 639-3). These represent the same language — this is a code-granularity correction, not a proxy substitution. _name_or_path is set to facebook/mms-tts-ful accordingly.

What this repository adds

facebook/mms-tts-* Hub checkpoints are inference-only releases that crash run_vits_finetuning.py. This repository applies three patches:

File	Change
config.json	pad_token_id set to 0 (was null)
tokenizer_config.json	pad_token entry added
preprocessor_config.json	Added — VitsFeatureExtractor config from ylacombe/mms-tts-eng-train

Model weights are not stored here. _name_or_path in config.json points to facebook/mms-tts-ful, so run_vits_finetuning.py loads weights from that checkpoint at training time.

preprocessor_config.json

Downloaded verbatim from ylacombe/mms-tts-eng-train. Values are VITS architecture constants shared by all MMS-TTS languages.

Field	Value
feature_extractor_type	VitsFeatureExtractor
feature_size	80
hop_length	256
max_wav_value	32768.0
n_fft	1024
padding_side	right
padding_value	0.0
return_attention_mask	False
sampling_rate	16000
spec_gain	1

Usage in finetune-hf-vits

{
  "model_name_or_path":     "rnjema-unima/mms-tts-pul-baseline",
  "feature_extractor_name": "rnjema-unima/mms-tts-pul-baseline",
  "dataset_name":           "google/WaxalNLP",
  "dataset_config_name":    "pul_tts",
  "audio_column_name":      "audio",
  "text_column_name":       "text",
  "train_split_name":       "train",
  "eval_split_name":        "validation"
}

Inference (after fine-tuning)

from transformers import VitsModel, VitsTokenizer
import torch, scipy

model     = VitsModel.from_pretrained("your-org/your-finetuned-model")
tokenizer = VitsTokenizer.from_pretrained("your-org/your-finetuned-model")

inputs = tokenizer("Your text in Pular.", return_tensors="pt")
with torch.no_grad():
    out = model(**inputs)

scipy.io.wavfile.write("output.wav", model.config.sampling_rate,
    out.waveform.squeeze().numpy())

Technical details

Architecture	VITS (end-to-end, no separate vocoder)
MMS match type	granularity_remap
pad_token_id	0
vocab_size	41
is_uroman	false
sampling_rate	16000 Hz

References

BibTex Citation:

This model was developed by Vineel Pratap et al. from Meta AI. If you use the model, consider citing the MMS paper:

@article{pratap2023mms, title={Scaling Speech Technology to 1,000+ Languages}, author={Vineel Pratap and Andros Tjandra and Bowen Shi and Paden Tomasello and Arun Babu and Sayani Kundu and Ali Elkahky and Zhaoheng Ni and Apoorv Vyas and Maryam Fazel-Zarandi and Alexei Baevski and Yossi Adi and Xiaohui Zhang and Wei-Ning Hsu and Alexis Conneau and Michael Auli}, journal={arXiv}, year={2023} }

Notable Acknowledgements:

• WAXAL paper — arXiv:2602.02734
• WAXAL dataset
• finetune-hf-vits
• facebook/mms-tts-ful
• preprocessor_config.json source — ylacombe/mms-tts-eng-train