A memory-efficient neural ODE framework based on high-level adjoint differentiation

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.