megatron.core.tensor_parallel.random#
Description
Classes
|
This function is adapted from torch.utils.checkpoint with two main changes: 1) torch.cuda.set_rng_state is replaced with _set_cuda_rng_state 2) the states in the model parallel tracker are also properly tracked/set/reset. |
Tracker for the cuda RNG states. |
Functions
|
Checkpoint a model or part of the model. |
Get cuda rng tracker. |
|
Initialize model parallel cuda seed. |