| language: ar | |
| license: cc-by-nc-sa-4.0 | |
| tags: | |
| - text-to-speech | |
| - arabic | |
| - f5-tts | |
| library_name: f5-tts | |
| pipeline_tag: text-to-speech | |
| # xttsv2 | |
| Mirror of [SWivid/Habibi-TTS](https://huggingface.co/SWivid/Habibi-TTS) — open-source Arabic dialectal TTS built on F5-TTS. | |
| Original work: *Habibi: Laying the Open-Source Foundation of Unified-Dialectal Arabic Speech Synthesis* — see https://github.com/SWivid/Habibi-TTS and https://arxiv.org/abs/2601.13802. | |
| ## Licensing (per-dialect, from upstream) | |
| | Dialect / Model | License | | |
| |---|---| | |
| | Unified, SAU, UAE | CC-BY-NC-SA-4.0 (non-commercial) | | |
| | MSA, EGY, IRQ, MAR, ALG | Apache 2.0 | | |
| If you intend commercial use, restrict yourself to the Apache-2.0 dialects (MSA / EGY / IRQ / MAR / ALG). | |
| ## Source attribution | |
| - Upstream model: `SWivid/Habibi-TTS` | |
| - Underlying TTS framework: `SWivid/F5-TTS` | |
| - Vocoder: `charactr/vocos-mel-24khz` | |