[T-A]

the Apertus model was trained on the Alps supercomputer, operational at CSCS since September 2024, a data center of over 10'000 top-of-the-line NVIDIA Grace-Hopper chips

https://log.alets.ch/110/