You can deploy this model and start earning money today!

alakxender/
mms-tts-div-ft-spk01-m01

Name: mms-tts-div-ft-spk01-m01
Availability: InStock

alakxender/voice-synthetic

Views

No views yet

Divehi TTS – Male Voice (VITS-based)

This is a fine-tuned VITS (Variational Inference with adversarial learning for end-to-end Text-to-Speech) model for Divehi speech synthesis. The model produces Male voice audio from Thaana-scripted Divehi text. Fine-tuned from Meta’s MMS-TTS architecture using a curated dataset of synthetic Divehi speech.

Model Details

Field	Value
Model ID	`alakxender/mms-tts-div-ft-spk01-m01`
Base Architecture	MMS-TTS (VITS)
Language	Divehi (dv)
Voice	Male
Sampling Rate	16 kHz
Tokenizer	VITSTokenizer
Inference Engine	Transformers (🤗 Hugging Face)

Usage

python

from transformers import VitsModel, VitsTokenizer
import torchaudio

tokenizer = VitsTokenizer.from_pretrained("alakxender/mms-tts-div-ft-spk01-m01")
model = VitsModel.from_pretrained("alakxender/mms-tts-div-ft-spk01-m01")

text = "މޫސުން ވަރަށް ގޯސްވެ، ފުވައްމުލަކުން ފެށިގެން އައްޑުއަށް އޮރެންޖް އެލާޓް ނެރެފި"
inputs = tokenizer(text, return_tensors="pt")
waveform = model.generate(**inputs).waveform[0]

torchaudio.save("output.wav", waveform.unsqueeze(0), 16000)

Evaluation Summary

Model: alakxender/mms-tts-div-ft-spk01-m01
Evaluated Samples: 3
Avg Estimated MOS (UTMOS): 3.151
json
1{ 2 "5": "Excellent (very natural)", 3 "4": "Good (mostly natural)", 4 "3": "Fair (some robotic quality)", 5 "2": "Poor (noticeably unnatural)", 6 "1": "Bad (unintelligible or very synthetic)" 7}
Artifacts:
- 🎵 Audio: outputs/audio/
- 📊 Spectrograms: outputs/spectrograms/
- 📄 Report: outputs/report.txt
- 📈 MOS Scores: outputs/mos_scores.txt

alakxender/mms-tts-div-ft-spk01-m01

Divehi TTS – Male Voice (VITS-based)

Model Details

Usage

Evaluation Summary

Acknowledgements

alakxender/
mms-tts-div-ft-spk01-m01