Expand description
CUDA runtime sanity-check helpers. See cuda::ensure_cuda_if_expected.
Backend device helpers for the Burn training stack.
After phase 5 of the Burn migration (#82), libtorch’s CUDA runtime
guard is gone — Burn picks its device from the backend type
parameter (B::Device) instead of through a runtime probe. This
module just exposes the default-device helper trainers use.
Functions§
- default_
burn_ device - Backend-generic default device helper for the Burn path.