Frank Lin 249e65b92d Graph-Safe RNG State Exchange for Tensor Parallelism (#114068) ... See #113541 The PR allows for registering and controlling multiple RNG states using indices, ensuring cudagraph-safe operations, and includes both C++ and Python API changes to support this functionality. cc @eellison @anijain2305 @jansel @ezyang @ptrblck @csarofeen @mcarilli Pull Request resolved: https://github.com/pytorch/pytorch/pull/114068 Approved by: https://github.com/ezyang, https://github.com/eqy, https://github.com/xuzhao9		2024-03-27 01:14:38 +00:00
..
benchmark
core	Graph-Safe RNG State Exchange for Tensor Parallelism (#114068)	2024-03-27 01:14:38 +00:00
cuda	Remove unused variables (#122496)	2024-03-22 18:04:09 +00:00
hip
macros	Remove C10_FALLTHROUGH (#120157)	2024-02-21 06:18:58 +00:00
mobile	Fix broken lint after #116876 (#122253)	2024-03-20 04:09:00 +00:00
test	Make SparseCsr a functionality dispatch key (#120703)	2024-03-01 13:28:46 +00:00
util	[Fix] Fixed behaviour for the conversion of complex tensors to bool (#121803)	2024-03-14 13:35:15 +00:00
xpu	Revert "Support gpu trace on XPU (#121795)"	2024-03-21 20:33:16 +00:00
BUCK.oss
BUILD.bazel
build.bzl
CMakeLists.txt	Remove redundant CMake NUMA code (#119650)	2024-02-12 21:53:44 +00:00
ovrsource_defs.bzl