strisunshinewentingwang/pytorch

Pull Requests Commits

Delete stable-sort-only-works-on-cpu warning (#61685)

malfet committed 3 years ago

e2c3049e

Compare DDP static graph (C++ core) with legacy DDP forward and backward delay. (#61507)

bowangbj committed 3 years ago

e098e900

Fix hardswish inplace op for strided tensor with skipped elements (#61622)

Akshit Khurana committed 3 years ago

7a3b05ea

det_backward: more robust and with complex support (#58195)

nikitaved committed 3 years ago

fce85480

[nnc] Added a new API to distribute loop and all its parents (#61293)

navahgar committed 3 years ago

bd360ebe

[nnc] Added a new API to compress all buffers in a given statement (#61087)

navahgar committed 3 years ago

76f09746

[nnc] Modified the semantics of reorder in using permutation (#61085)

navahgar committed 3 years ago

2908d3eb

Revert [DDP] Support not all outputs used in loss calculation (#61497)

rohan-varma committed 3 years ago

71775093

Revert [DDP] Support for multiple backwards (#61401)

rohan-varma committed 3 years ago

25f9c35d

Back out "[DDP] Disable reducer hooks from running outside of DDP backwards." (#61399)

rohan-varma committed 3 years ago

38ac9e69

Revert D29701479: [pytorch][PR] Remove `_broadcast_object()` from `ZeroRedundancyOptimizer`

yuguo68 committed 3 years ago

a50a389c

Fix for get_buffer(): check buffers by name instead of value (#61429)

jbschlosser committed 3 years ago

aa01a7a6

CopyBackward: Remove redundant src_device and unnecessary copy=True (#60025)

peterbell10 committed 3 years ago

54071085

Add .github for CODEOWNERS (#61598)

seemethere committed 3 years ago

da667e2d

changed launch bounds for upsample_linear1d fwd, bwd from 1024 to 512 (#61307)

Fuzzkatt committed 3 years ago

8afb65b6

Register Saved Tensors hooks (#60663)

Varal7 committed 3 years ago

ee5a97de

[static runtime] Use at::allclose to test NNC sigmoid (#61566)

d1jang committed 3 years ago

94965212

Remove `_broadcast_object()` from `ZeroRedundancyOptimizer` (#61539)

awgu committed 3 years ago

9b5d9b40

[pytorch][profiler] Fix division by 0 in computeFlops (#61676)

valentinandrei committed 3 years ago

e3d5619f

Avoid redefining `__BYTE_ORDER` (#60346)

Dimitrije Jankov committed 3 years ago

70e94bb1

Grammatical update of tech docs (#61547)

JINAY08 committed 3 years ago

a9c35800

add trainer hook functions (#60785)

gcramer23 committed 3 years ago

5a5c7f56

refactor ps benchmark (#60784)

gcramer23 committed 3 years ago

304c02ee

Release GIL as much as possible in dist_autograd pybind. (#61593)

pritamdamania committed 3 years ago

7d2ea9a8

Avoid holding GIL while calling retrieveContext. (#61588)

pritamdamania committed 3 years ago

5ebc7c9f

[Metal] Do not use read/write textures in concat shaders (#61074)

SS-JIA committed 3 years ago

f2adbff3

Skip Bfloat16 support when building for VSX (#61630)

malfet committed 3 years ago

80bdfd64

[TensorExpr] Do not fuse float16 values. (#61569)

Mikhail Zolotukhin committed 3 years ago

43a2f7c2

Make broadcast_object_list accept a device parameter. (#61305)

bowangbj committed 3 years ago

ab273995

[pruner] fix activation handles logic (#61592)

kazhou committed 3 years ago

9b3cbeaf

Newer Older