strisunshinewentingwang/pytorch

Pull Requests Commits

[LTC] Make ComputePostOrder accept const T pointers (#88773)

alanwaketan committed 2 years ago

3b8245ab

[Inductor] Build FX Linear + Permute Vertical Fusion in Inductor (#88566)

jiawenliu64 committed 2 years ago

48b58930

Revert "Symintify `broadcast_to` (#88776)"

pytorchmergebot committed 2 years ago

d157fca5

[FSDP][Perf] Do not call `pad` in no-padding case (#88769)

awgu committed 2 years ago

6bf2776a

[dynamo] `VariableTracker.call_method` requires a name (#88311)

bertmaher committed 2 years ago

d3178465

[nnc] Disable opaque pointers mode in LLVM backend to allow getPointerElementType (#88798)

bertmaher committed 2 years ago

1e4079a4

Change TORCH_INTERNAL_ASSERT to TORCH_CHECK and add a nice error message (#88804)

PanosAntoniadis committed 2 years ago

656d0de6

Switch to setup-nvidia action (#88757)

huydhn committed 2 years ago

79b049af

Make TorchElastic timer importable on Windows (#88522)

malfet committed 2 years ago

f98edfcc

Symintify `adaptive_avg_pool3d` (#88783)

nkaretnikov committed 2 years ago

4b898a73

Symintify `broadcast_to` (#88776)

nkaretnikov committed 2 years ago

3a09d9a1

add DisableTorchFunction that matches DisableTorchDispatch (#88219)

samdow committed 2 years ago

c0ecce15

rename DisableTorchFunction to DisableTorchFunctionSubclass (#88218)

samdow committed 2 years ago

7f28be10

torchdynamo: add convolution add(relu) inplace fusion kernel (#88048)

XiaobingSuper committed 2 years ago

3e43ff27

add hack to allow hybrid compressed sparse comparison in assertEqual (#88749)

pmeier committed 2 years ago

e6561291

[MPS] Fix softplus (#88555)

qqaatw committed 2 years ago

7c353eb3

Support src_mask and src_key_padding_mask for Better Transformer (#88488)

sgrigory committed 2 years ago

7ad87f63

[caffe2][tourch] Optimize BatchBoxCox (#87585)

efiks committed 2 years ago

dcefea27

[vision hash update] update the pinned vision hash (#88742)

pytorchmergebot committed 2 years ago

e87c79ca

[dynamo] fixes dict changed during runtime error (#87526)

anijain2305 committed 2 years ago

cf04b36c

Do not flag models in dashboard due to NaN values (#88792)

williamwen42 committed 2 years ago

0b8889c7

Add absolute latency to dashboard (#88790)

williamwen42 committed 2 years ago

6e3555ed

add stride constraints to fallbacks (#88534)

eellison committed 2 years ago

23815480

[cuDNN][cuDNN V8 API] Match V7 API behavior for `channels_last` stride coercion for cuDNN (#88699)

eqy committed 2 years ago

fb5c6ae6

disable test that times out in fbcode (#88758)

mikey dagitses committed 2 years ago

59115e61

Fix dynamo dashboard passrate denominator (#88777)

williamwen42 committed 2 years ago

16bd3638

Fix Docker image generation (#88741)

malfet committed 2 years ago

4f18739b

[Dynamo] Fix Tensor.T trace (#88642)

Akshit Khurana committed 2 years ago

7006ac6e

Revert "[3/n] Thread PG: add threaded PG implementation (#88627)"

pytorchmergebot committed 2 years ago

c7fc7104

[ao] qconfig.py fix public v private (#87515)

HDCharles committed 2 years ago

6fe4ccc7

Newer Older