Instructions to use zenlm/zen-translator with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use zenlm/zen-translator with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "translation" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # 'pip install "transformers<5.0.0' from transformers import pipeline pipe = pipeline("translation", model="zenlm/zen-translator")# Load model directly from transformers import ZenTranslatorForSpeechTranslation model = ZenTranslatorForSpeechTranslation.from_pretrained("zenlm/zen-translator", dtype="auto") - Notebooks
- Google Colab
- Kaggle
| """ | |
| Zen Translator - Real-time multimodal translation with lip sync and voice cloning. | |
| Built on: | |
| - Qwen3-Omni: Real-time speech understanding and translation | |
| - CosyVoice 2.0: Ultra-low latency voice cloning (150ms) | |
| - Wav2Lip: Accurate lip synchronization | |
| Features: | |
| - 18 input languages, 10 output languages | |
| - News anchor voice finetuning for accurate translation | |
| - Sub-second end-to-end latency | |
| - WebRTC streaming support | |
| """ | |
| __version__ = "0.1.0" | |
| __author__ = "Hanzo AI / Zen LM" | |
| from .config import TranslatorConfig | |
| from .pipeline import TranslationPipeline | |
| __all__ = [ | |
| "TranslationPipeline", | |
| "TranslatorConfig", | |
| "__version__", | |
| ] | |