EGTTS V0.1

🚀 EGTTS V0.1

EGTTS V0.1 is a state - of - the - art text - to - speech (TTS) model tailored for Egyptian Arabic. Based on the XTTS v2 architecture, it can convert written Egyptian Arabic text into natural - sounding speech, facilitating seamless communication in diverse applications like voice assistants, educational tools, and chatbots.

🚀 Quick Start

📦 Installation

Dependencies to install

pip install git+https://github.com/coqui-ai/TTS

pip install transformers

pip install deepspeed

💻 Usage Examples

Load the model

import os
import torch
import torchaudio
from TTS.tts.configs.xtts_config import XttsConfig
from TTS.tts.models.xtts import Xtts

CONFIG_FILE_PATH = 'path/to/config.json'
VOCAB_FILE_PATH = 'path/to/vocab.json'
MODEL_PATH = 'path/to/model'
SPEAKER_AUDIO_PATH = 'path/to/speaker.wav'

print("Loading model...")
config = XttsConfig()
config.load_json(CONFIG_FILE_PATH)
model = Xtts.init_from_config(config)
model.load_checkpoint(config, checkpoint_dir=MODEL_PATH, use_deepspeed=True, vocab_path=VOCAB_FILE_PATH)
model.cuda()

print("Computing speaker latents...")
gpt_cond_latent, speaker_embedding = model.get_conditioning_latents(audio_path=[SPEAKER_AUDIO_PATH])

Run the model

from IPython.display import Audio, display

text = "صباح الخير"
print("Inference...")
out = model.inference(
    text,
    "ar",
    gpt_cond_latent,
    speaker_embedding,
    temperature=0.75,
)

AUDIO_OUTPUT_PATH = "path/to/output_audio.wav"
torchaudio.save("xtts_audio.wav", torch.tensor(out["wav"]).unsqueeze(0), 24000)
display(Audio(AUDIO_OUTPUT_PATH, autoplay=True))

✨ Features

• Widget Information: | Property | Details | |----------|---------| | Source Audio | sample.flac | | Output Text | صباح الخير. | | License | other | | License Name | coqui - public - model - license | | Language | ar | | Base Model | coqui/XTTS - v2 | | Pipeline Tag | text - to - speech |

🔗 Try It Out

✨ Experience the magic of EGTTS V0.1 live! Try the model directly through this HuggingFace Space.

🔍 Explore the Code

💻 Dive into the implementation! Check out the full code on GitHub.

📄 License

The model uses the coqui - public - model - license license.

📚 Citation

@misc{OmarSamir,
      author = {Omar Samir, Youssef Waleed, Youssef Tamer ,and Amir Mohamed},
      title = {Fine-Tuning XTTS V2 for Egyptian Arabic},
      year = {2024},
      url = {https://github.com/joejoe03/Egyptian-Text-To-Speech},
}