master
[FEATURE] Add tanh approximation for GeLU activation (#21034)
[master] Node elimination graph pass (#21046)
[BUGFIX] Reenable fwd conv engine 5 on test_group_conv2d_16c (#21104)
[BUGFIX] Fix threadsafety and shutdown issues with threaded_engine_perdevice (#21110)
Port top-level-project updates from v1.x branch (#21162)
[master][clang-format] Re-format cc. .h. .cu files; cond. (#20704)
[Master] Clang-formatter: only src/ directory (#20571)
MXNet FFI for Operator Imperative Invocation (#17510)
Refactor SupportDNNL functions (#21032)
Improve bf16 support (#21002)
[v2.0][LICENSE] Port #20493 (#20608)
Resolve the conflict with PR#20499 (#20887)
Improve AMP, bf16 support. Support oneDNN ops in AMP (#20753)
Avoid modifying loaded library map while iterating in lib_close() (#20941)
Diversify default RNG seed (#21058)