Upload README.md with huggingface_hub

README.md

@@ -1,27 +1,33 @@
-# Luganda TTS (lug-tts-v4)
+---
+language: lg
+tags:
+- tts
+- text-to-speech
+- luganda
+- nemo
+- fastpitch
+- hifigan
+license: apache-2.0
+---
 
-Luganda Text-to-Speech model trained with NVIDIA NeMo.
+# Luganda TTS v4
+
+Text-to-Speech model for Luganda language using NVIDIA NeMo FastPitch + HiFi-GAN.
 
 ## Models
-- `luganda_fastpitch_final.nemo` - FastPitch spectrogram generator  
-- `luganda_hifigan_final.nemo` - HiFi-GAN vocoder
+
+- **luganda_fastpitch_final.nemo**: FastPitch mel-spectrogram generator fine-tuned on Luganda
+- **tts_en_hifigan.nemo**: HiFi-GAN vocoder (pretrained English)
+- **pitch_stats.json**: Luganda pitch normalization statistics
 
 ## Usage
-```python
-from nemo.collections.tts.models import FastPitchModel, HifiGanModel
-import soundfile as sf
 
-fastpitch = FastPitchModel.restore_from("luganda_fastpitch_final.nemo")
-hifigan = HifiGanModel.restore_from("luganda_hifigan_final.nemo")
 
-text = "Oli otya?"
-parsed = fastpitch.parse(text)
-spec = fastpitch.generate_spectrogram(tokens=parsed)
-audio = hifigan.convert_spectrogram_to_audio(spec=spec)
-sf.write("output.wav", audio.cpu().numpy()[0], 22050)
-```
 
 ## Training
-- Dataset: Sunbird/salt Luganda studio recordings
-- Steps: 5,000 (FastPitch fine-tuning)
-- Test Results: 8/8 phrases passed
+
+Fine-tuned from English FastPitch on ~2000 Luganda samples from Sunbird/salt dataset.
+
+## License
+
+Apache 2.0