Nari Labs

Nari Labs is a South Korean startup. They work on making text-to-speech technology better. Their main product is Dia. Dia is an open-source text-to-speech model. It has 1.6 billion parameters. This model turns text into ultra-realistic dialogue. It sounds like a real conversation. Nari Labs Dia can be used by researchers, developers, and content creators. They can use it for virtual assistants, gaming, audiobooks, and accessibility tools. Benefits Nari Labs Dia has several key advantages. It can generate realistic dialogue. It has customizable speaker tones, emotional inflections, and nonverbal cues like laughter and sighs. The model supports zero-shot voice cloning. It can mimic a voice from just a few seconds of reference audio. It performs in real-time on a single GPU. This makes it efficient and accessible. Nari Labs Dia is open-source. Anyone can use, modify, and improve the technology. Use Cases Nari Labs Dia can be used in many settings. It is perfect for creating podcasts, audiobooks, and game characters. The model can also be used in interactive applications and virtual assistants. It makes them more engaging and lifelike. Its ability to generate realistic dialogue makes it an excellent tool. Content creators and developers can use it to enhance their audio projects. Additional Information Nari Labs Dia was developed by two undergraduate students. Toby Kim was one of them. They did not have any external funding. They used Google''s TPU Research Cloud and Hugging Face''s ZeroGPU grant. They created this innovative technology. The model is available on GitHub and Hugging Face. Users can access pretrained checkpoints, inference code, and a demo for easy testing. The development team is working on new features and improvements. Nari Labs Dia is an exciting tool for the future of AI audio generation.