strisunshinewentingwang/pytorch

Pull Requests Commits

[xnnpack][on-device] executor class (#88778)

mcr229 committed 1 year ago

d5e1e2f0

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

pytorchmergebot committed 1 year ago

29550e2c

[DataPipe] Deprecating drop_empty_batches from Filter and other functional APIs (#88693)

ejguan committed 1 year ago

90cf14dd

Bring Unfold/Fold param doc order in line with code (#88819)

felixdivo committed 1 year ago

98ecd065

[14/N] Refactor _new_process_group_helper() to remove repeated code (#88351)

H-Huang committed 1 year ago

1d54ce9d

Add warnings & regressions info text (#88837)

williamwen42 committed 1 year ago

4bcf2c53

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

alanwaketan committed 1 year ago

3b8245ab

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

jiawenliu64 committed 1 year ago

48b58930

Revert "Symintify `broadcast_to` (#88776)"

pytorchmergebot committed 1 year ago

d157fca5

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

awgu committed 1 year ago

6bf2776a

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

bertmaher committed 1 year ago

d3178465

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

bertmaher committed 1 year ago

1e4079a4

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

PanosAntoniadis committed 1 year ago

656d0de6

Switch to setup-nvidia action (#88757)

huydhn committed 1 year ago

79b049af

Make TorchElastic timer importable on Windows (#88522)

malfet committed 1 year ago

f98edfcc

Symintify `adaptive_avg_pool3d` (#88783)

nkaretnikov committed 1 year ago

4b898a73

Symintify `broadcast_to` (#88776)

nkaretnikov committed 1 year ago

3a09d9a1

add DisableTorchFunction that matches DisableTorchDispatch (#88219)

samdow committed 1 year ago

c0ecce15

rename DisableTorchFunction to DisableTorchFunctionSubclass (#88218)

samdow committed 1 year ago

7f28be10

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

XiaobingSuper committed 1 year ago

3e43ff27

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

pmeier committed 1 year ago

e6561291

[MPS] Fix softplus (#88555)

qqaatw committed 1 year ago

7c353eb3

Support src_mask and src_key_padding_mask for Better Transformer (#88488)

sgrigory committed 1 year ago

7ad87f63

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

efiks committed 1 year ago

dcefea27

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

pytorchmergebot committed 1 year ago

e87c79ca

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

anijain2305 committed 1 year ago

cf04b36c

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

williamwen42 committed 1 year ago

0b8889c7

Add absolute latency to dashboard (#88790)

williamwen42 committed 1 year ago

6e3555ed

add stride constraints to fallbacks (#88534)

eellison committed 1 year ago

23815480

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

eqy committed 1 year ago

fb5c6ae6

Newer Older