Hugging Face's logo

hamdallah
/
Sofelia-TTS

Text-to-Speech
Transformers
Safetensors
Arabic
English
qwen2
text-generation
tts
audio
speech
palestinian-arabic
arabic
voice-cloning
miratts
sofelia
text-generation-inference
Model card Files
xet
Community
1

πŸ‡΅πŸ‡Έ Sofelia-TTS πŸ‡΅πŸ‡Έ

Palestinian Arabic Text-to-Speech Model

Palestine will be free πŸ•ŠοΈ

🌟 Model Description

Sofelia-TTS is a fine-tuned Text-to-Speech (TTS) model specifically trained for Palestinian Arabic dialect. This model brings the beautiful sounds of Palestinian speech to AI, preserving and celebrating the linguistic heritage of Palestine.

Built on top of YatharthS/MiraTTS, Sofelia-TTS captures the unique phonetic characteristics, intonation patterns, and prosody of Palestinian Arabic, making it ideal for:

  • πŸŽ™οΈ Voice cloning with Palestinian Arabic speech
  • πŸ“š Audiobook generation in Palestinian dialect
  • πŸ—£οΈ Virtual assistants that speak authentic Palestinian Arabic
  • πŸŽ“ Educational tools for learning and preserving the Palestinian dialect
  • 🎬 Content creation for Palestinian media and storytelling

Dedicated to Palestine: This model is a tribute to the resilience, culture, and spirit of the Palestinian people. May their voices be heard loud and clear across the world. πŸ‡΅πŸ‡Έ

🎯 Key Features

  • βœ… High-quality voice cloning: Clone any voice with just a few seconds of reference audio
  • βœ… Palestinian Arabic dialect: Authentic pronunciation and intonation
  • βœ… Fast inference: Optimized for real-time generation
  • βœ… Flexible context: Supports variable-length reference audio
  • βœ… Open source: Free to use and improve

πŸ“Š Model Details

Attribute Value
Model Type Text-to-Speech (TTS)
Base Model YatharthS/MiraTTS
Architecture Transformer-based Language Model + Audio Codec
Training Language Palestinian Arabic (ar-PS)
Dataset Private Dataset
Sample Rate 16,000 Hz
License Apache 2.0
Model Size ~0.6B parameters
Precision BF16/FP32
Framework PyTorch + Transformers

πŸš€ Quick Start

Installation 

# Install required packages
uv pip install git+https://github.com/ysharma3501/MiraTTS.git

Usage (Python) 

from mira.model import MiraTTS
from IPython.display import Audio
mira_tts = MiraTTS('hamdallah/Sofelia-TTS') ## downloads model from huggingface

file = "reference_file.wav" ## can be mp3/wav/ogg or anything that librosa supports
text = "Ω…Ψ±Ψ­Ψ¨Ψ§ΨŒ ΩƒΩŠΩ Ψ§Ω„Ψ­Ψ§Ω„ΨŸ Ω‡Ψ°Ψ§ Ω†Ω…ΩˆΨ°Ψ¬ Ω„Ω„Ω‡Ψ¬Ψ© Ψ§Ω„ΩΩ„Ψ³Ψ·ΩŠΩ†ΩŠΨ©."

context_tokens = mira_tts.encode_audio(file)
audio = mira_tts.generate(text, context_tokens)

Audio(audio, rate=48000)

🎀 Example Prompts

Try these Palestinian Arabic phrases:

# Greetings
"Ω…Ψ±Ψ­Ψ¨Ψ§ΨŒ ΩƒΩŠΩ Ψ­Ψ§Ω„ΩƒΨŸ"  # Hello, how are you?
"Ψ£Ω‡Ω„Ψ§ ΩˆΨ³Ω‡Ω„Ψ§ ΩΩŠΩƒ"      # Welcome

# Common expressions
"يا Ψ³Ω„Ψ§Ω…ΨŒ Ω‡Ψ°Ψ§ Ψ±Ψ§Ψ¦ΨΉ"   # Wow, this is amazing
"Ω…Ψ§ Ψ΄Ψ§Ψ‘ Ψ§Ω„Ω„Ω‡"         # Mashallah
"Ψ§Ω„Ω„Ω‡ ΩŠΨΉΨ·ΩŠΩƒ Ψ§Ω„ΨΉΨ§ΩΩŠΨ©"  # God give you wellness

# About Palestine
"ΩΩ„Ψ³Ψ·ΩŠΩ† Ψ­Ψ±Ψ© ΨΉΩ„Ω‰ Ψ·ΩˆΩ„"  # Palestine is free for ever
"Ψ§Ω„Ω‚Ψ―Ψ³ ΨΉΨ§Ψ΅Ω…Ψ© ΩΩ„Ψ³Ψ·ΩŠΩ† Ψ§Ω„Ψ£Ψ¨Ψ―ΩŠΨ©"     # Jerusalem is the eternal capital of Palestine
"Ψ³Ω†ΨΉΩˆΨ― ΩŠΩˆΩ…Ψ§Ω‹ Ψ₯Ω„Ω‰ Ψ―ΩŠΨ§Ψ±Ω†Ψ§"         # We will return one day to our homes

πŸŽ“ Training Details

Training Data

  • Dataset: 400 Hours Palestinian Speech
  • Language: Palestinian Arabic dialect
  • Hours of audio: High-quality Palestinian speech recordings
  • Preprocessing: Audio normalized and resampled to 16kHz

Training Configuration

Hyperparameter Value
Learning Rate 2e-4 (initial), 1e-5 (refinement)
Batch Size 8 (effective: 2 per device Γ— 4 accumulation steps)
Training Steps 2000+
Warmup Steps 100
Max Audio Length 20-30 seconds
Optimizer AdamW
LR Scheduler Cosine with warmup
Gradient Clipping 1.0
Precision BF16 (H100) / FP32
Hardware NVIDIA H100 / A100 GPU

Training Process

The model was trained using a two-phase approach:

  1. Foundation Phase: High learning rate (2e-4) for initial adaptation to Palestinian Arabic
  2. Refinement Phase: Lower learning rate (1e-5) with NEFTune noise for stability and quality

πŸ“ˆ Model Performance

The model achieves:

  • βœ… Natural prosody matching Palestinian Arabic speech patterns
  • βœ… Clear pronunciation of Arabic phonemes
  • βœ… Voice similarity to reference audio
  • βœ… Stable generation without artifacts or repetitions
  • βœ… Fast inference suitable for real-time applications

πŸ› οΈ Advanced Usage

Running the model using batching 

file = "reference_file.wav" ## can be mp3/wav/ogg or anything that librosa supports
text = ["Ω…Ψ±Ψ­Ψ¨Ψ§ΨŒ ΩƒΩŠΩ Ψ­Ψ§Ω„ΩƒΨŸ", "Ψ¨Ψͺعرف Ψ₯Ω†Ω‡ Ψ§Ω†Ψ§ Ψ¨Ω‚Ψ―Ψ± Ψ§Ψ­ΩƒΩŠ ΩΩ„Ψ³Ψ·ΩŠΩ†ΩŠ و English Ω…ΨΉ Ψ¨ΨΉΨΆ Without Errors."]

context_tokens = [mira_tts.encode_audio(file)]

audio = mira_tts.batch_generate(text, context_tokens)

Audio(audio, rate=48000)

Adjusting Generation Parameters 

# More creative/variable output

mira_tts.set_params(
    top_p=0.95, 
    top_k=20, 
    temperature=0.01, # Higher = more variation
    max_new_tokens=1024, 
    repetition_penalty=2.2, 
    min_p=0.05
)

πŸ’‘ Tips for Best Results

  1. Reference Audio Quality:

    • Use clean audio without background noise
    • 3-10 seconds of speech is ideal
    • Ensure audio is 16kHz sample rate

  2. Text Input:

    • Use proper Arabic script (not Arabizi/transliteration)
    • Palestinian dialect works best
    • Avoid very long sentences (split into shorter segments)

  3. Generation Parameters:

    • temperature=0.7: Good default for natural speech
    • temperature=0.5: More stable, less variation
    • temperature=0.9: More expressive, more variation

🌍 About Palestinian Arabic

Palestinian Arabic is a Levantine Arabic dialect spoken by the Palestinian people. It has unique characteristics:

  • Phonology: Preservation of Classical Arabic /q/ as glottal stop [Κ”]
  • Vocabulary: Rich in Levantine and unique Palestinian terms
  • Intonation: Distinctive melodic patterns
  • Regional Variants: Urban (Jerusalem, Hebron) vs. Rural vs. Bedouin varieties

This model captures these linguistic features, making it authentic and representative of Palestinian speech.

πŸ‡΅πŸ‡Έ Message of Solidarity

This model is dedicated to the Palestinian people and their enduring struggle for freedom, dignity, and justice. Through technology, we preserve and celebrate Palestinian culture, language, and identity.

Free Palestine πŸ‡΅πŸ‡Έ

"We will not be erased. Our voices will echo through time, in every language model, every algorithm, every line of code. Palestine lives, and so does its voice."

πŸ“œ License

This model is released under the Apache 2.0 License, making it free for:

  • βœ… Commercial use
  • βœ… Modification and distribution
  • βœ… Private use
  • βœ… Patent use

πŸ™ Acknowledgments

  • Base Model: YatharthS/MiraTTS - Thank you for the excellent foundation
  • Dataset: Palestinian Arabic speakers who contributed their voices
  • Community: The open-source AI community for tools and support
  • Palestine: For being the inspiration and purpose behind this work

πŸ“ž Contact & Support

πŸ”— Related Resources

πŸ“š Citation

If you use this model in your research or projects, please cite:

@misc{sofelia-tts-2026,
  author = {Hamdallah},
  title = {Sofelia-TTS: Palestinian Arabic Text-to-Speech Model},
  year = {2026},
  publisher = {Hugging Face},
  journal = {Hugging Face Model Hub},
  howpublished = {\url{https://huggingface.co/hamdallah/Sofelia-TTS}},
}

πŸ‡΅πŸ‡Έ FREE PALESTINE πŸ‡΅πŸ‡Έ

Ψͺحيا ΩΩ„Ψ³Ψ·ΩŠΩ† Ψ­Ψ±Ψ© أبية

Long Live Free Palestine

πŸ•ŠοΈ ✊ πŸ‡΅πŸ‡Έ

Made with ❀️ for Palestine

Downloads last month
46
Safetensors
Model size
0.5B params
Tensor type
F32
Β·
Inference Providers NEW
Text-to-Speech
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for hamdallah/Sofelia-TTS

Base model

YatharthS/MiraTTS
Finetuned
(8)
this model
Quantizations
1 model