Add support for gradient accumulation (#46)

Remove use of .get() in training.cpp.

Add gradient accumulation test for supervised learning.

Remove loss normalization by grad_accumulation_steps for supervised t…

adding gradient accumulation

Better handling of g++ compiler requirement in CMake build. (#44)

Add support for gradient accumulation. Initial support for supervised…

Update CMakeLists.txt.

Better handling of g++ compiler requirement in CMake build.

Add tensor continuity checks in allreduce and broadcast methods. (#42)

Add check for tensor continuity in broadcast method.

Add checks for tensor continuity in allreduce methods.

Add checks for tensor continuity in allreduce methods.

Improve advantage calculation (#41)

Apply per-rank seeding to rollout and replay buffer sampling (#40)

Tkurth/multi env 2 (#37)

Disambiguate module locations for Fortran simulation example to fix r…

Disambiguate module locations for Fortran simulation example to fix r…

Disambiguate module locations for Fortran simulation example to fix r…

tkurth/ppo multi env (#32)

fixing CHECK_MPI according to issue #30 (#31)

Add more robust check for MPI_Comm_f2c during build. (#29)

Add more robust check for MPI_Comm_f2c during build.

Add training/inference functions that support multiple input/output a…