| --- |
| license: cc-by-nc-sa-4.0 |
| library_name: habibi-tts |
| datasets: |
| - sqrk/mixat-tri |
| - facebook/omnilingual-asr-corpus |
| - ArabicSpeech/MGB-3 |
| - MightyStudent/Egyptian-ASR-MGB-3 |
| - google/fleurs |
| - ArabicSpeech/MGB-5 |
| - Nexdata/UAE_Arabic_Spontaneous_Speech_Data |
| - AhmedBadawy11/UAE_100K |
| - adiren7/darija_speech_to_text |
| - KandirResearch/DarijaTTS-clean |
| - nadsoft/Jordan-Audio |
| language: |
| - ar |
| base_model: |
| - SWivid/F5-TTS |
| --- |
| |
| # A suite of specialized and unified dialectal Arabic TTS models |
|
|
| Paper: "Habibi: Laying the Open-Source Foundation of Unified-Dialectal Arabic Speech Synthesis" |
| Github: https://github.com/SWivid/Habibi-TTS |
|
|
|
|
| > [!NOTE] |
| > The unified, SAU, and UAE models are licensed under CC-BY-NC-SA-4.0, restricted by [SADA](https://www.kaggle.com/datasets/sdaiancai/sada2022) and [Mixat](https://huggingface.co/datasets/sqrk/mixat-tri). |
| > All the rest (ALG, EGY, IRQ, MAR, MSA) are released under Apache 2.0 license. |
