Views
No views yet
| Code | Dialect |
|---|---|
| MSA | Modern Standard Arabic (فصحى) |
| EGY | Egyptian (مصري) |
| SAU | Saudi (سعودي) |
| UAE | Emirati (إماراتي) |
| IRQ | Iraqi (عراقي) |
| MAR | Moroccan (مغربي) |
| ALG | Algerian (جزائري) |
1from f5_tts_mlx.cfm import F5TTS
2
3# Load directly from this repo
4f5tts = F5TTS.from_pretrained("pmax-agency/habibi-tts-mlx")
5
6# Generate Arabic speech
7wave, _ = f5tts.sample(
8 ref_audio,
9 text=["أهلاً وسهلاً، كيف حالك اليوم؟"],
10 duration=duration,
11 steps=8,
12)1pip install f5-tts-mlx
2python -m f5_tts_mlx.generate \
3 --model pmax-agency/habibi-tts-mlx \
4 --text "أهلاً وسهلاً" \
5 --output arabic_speech.wavema_model. prefixinitted, step)convert_weights=True conversionmodel_v1.safetensors file contains weights in PyTorch key format — f5-tts-mlx's from_pretrained handles the final key renaming and conv weight transposition automatically.