Apertus: Open language model for research & customization. Inspect training process, adapt to your needs. No download, requires account login. Explore Apertus now!
Apertus is an open language model designed for researchers, professionals, and enthusiasts to customize and experiment with. Unlike models offering limited access, Apertus provides full transparency, enabling users to inspect the entire training process. Trained on a massive dataset of 15 trillion tokens across over 1,000 languages (40% non-English), including underrepresented languages like Swiss German and Romansh, Apertus offers unique capabilities. Users can access the model through Swisscom or download it from Hugging Face for deployment in their projects. Available in two sizes (8 billion and 70 billion parameters), both models are released under a permissive open-source license, facilitating educational, research, societal, and commercial applications.
