habash2125/oromo_tts

# Arabic → Oromo Translator & TTS (Streamlit) Web app that takes **Arabic** text, translates it to **Oromo (Afaan Oromoo)**, and speaks the Oromo translation using Facebook MMS-TTS. - **Translation:** NLLB-200 distilled 600M (Arabic → Oromo) - **Speech:** MMS-TTS Oromo (VITS) --- ## Deploy on GitHub + Streamlit Community Cloud 1. **Push this project to GitHub** ```bash cd /path/to/oromo_tts git init git add . git commit -m "Arabic → Oromo translator & TTS" git branch -M main git remote add origin github.com git push -u origin main ``` (Create the repo first at github.com and replace `YOUR_USERNAME` with your GitHub username.) 2. **Deploy the app online (free)** - Go to share.streamlit.io. - Sign in with GitHub. - Click **New app**, choose your **oromo_tts** repo. - Set **Main file path** to `app.py`. - Click **Deploy**. The first run will download the models and may take a few minutes. Your app will get a public URL like `your-app-name.streamlit.app`. **Note:** Streamlit Cloud has memory limits. If the app runs out of memory, try Hugging Face Spaces (e.g. CPU Space with Streamlit SDK) and use the same `app.py` and `requirements.txt`. --- ## Run locally ### Setup ```bash git clone github.com cd oromo_tts python -m venv venv source venv/bin/activate # Windows: venv\Scripts\activate pip install -r requirements.txt ``` ### Run ```bash streamlit run app.py ``` Open the URL shown in the terminal (default: localhost). ## Usage 1. Enter or paste **Arabic** text in the text area. 2. Click **Translate & speak**. 3. View the **Oromo translation**, play the audio in the browser, or download the WAV file. Models are loaded once and cached. Adjust **TTS seed** to change prosody. ## License - MMS-TTS: **CC-BY-NC 4.0**. See model card. - NLLB-200: **CC-BY-NC 4.0**.