F5-TTS Romanian
First F5-TTS model fine-tuned for Romanian language.
Training Data
- Common Voice 17.0 Romanian: 35,289 examples
- TTS-Romanian (datadriven-company): 50,000 examples
- Total: ~85,000 examples, 173 hours of Romanian audio
Usage
pip install f5-tts
f5-tts_infer-cli \
--model F5TTS_v1_Base \
--ckpt_file model_last.pt \
--vocab_file vocab.txt \
--ref_audio your_voice.wav \
--ref_text "textul exact din clipul de referinta" \
--gen_text "textul pe care vrei sa il generezi" \
--output_file output.wav
Model Details
- Base model: F5TTS_v1_Base
- Fine-tuned for: Romanian language
- Tokenizer: character-based
- Training: 8 epochs
- Downloads last month
- 58