strisunshinewentingwang/pytorch

Pull Requests Commits

Windows nvtx for CUDA 11.6 require _wgetenv decl (#79798)

George Gensure committed 2 years ago

7d17e3b8

Revert "Fix BytesWarning in torch.load() (#74813)"

pytorchmergebot committed 2 years ago

e10cbe38

[ONNX] Refactor heavy memory usage tests

BowenBao committed 2 years ago

73f6601c

MPS: Add amax and amin Ops with tests (#79682)

abhudev committed 2 years ago

074dc746

Revert "Create a new Ubunutu 22.04 (jammy) build for platform010 (#77591)"

pytorchmergebot committed 2 years ago

dee43798

SherlockNoMad committed 2 years ago

b7888cbf

Do not bind `c10::irange` output by reference (#79815)

malfet committed 2 years ago

399b3dca

Fix BytesWarning in torch.load() (#74813)

aphedges committed 2 years ago

6c2e8119

Exploit symmetry in comparison operators to reduce no. of kernels

peterbell10 committed 2 years ago

706fa49c

disable modes during constructor

samdow committed 2 years ago

24243659

[primTorch] refs: nn.functional.{hard|soft}shrink (#78427)

kshitij12345 committed 2 years ago

cd3a0471

Revert "Back out "[const_fold] Set requires_grad based on the folded tensor; add device_for_folding option" (#79655)"

pytorchmergebot committed 2 years ago

e627dd54

[shard] make state_dict hook be consistent

wanchaol committed 2 years ago

bef2fecb

Add test for torchscripting nn.TransformerEncoder, including fast path (#79796)

erichan1 committed 2 years ago

06274d7a

[AMP] Use generic autocast in example, specify dtype (#79579)

eqy committed 2 years ago

eff74ed7

include cudnn_batch_norm, filter out view

SherlockNoMad committed 2 years ago

64ffc6f8

Create a new Ubunutu 22.04 (jammy) build for platform010 (#77591)

mehtanirav committed 2 years ago

71d82917

filter partitions

SherlockNoMad committed 2 years ago

cf9eab4a

started using mode-based tracing

Chillee committed 2 years ago

f5d7e5a1

[JIT] Add basic aliasing checks for tensor inputs

goldenxuett committed 2 years ago

1432a3d6

include get_item in the fusion group

SherlockNoMad committed 2 years ago

32313161

Add symmetric version of gpu_kernel_with_scalars

peterbell10 committed 2 years ago

e8727994

Add option for allowing non-fake inputs, add deepcopy impl

eellison committed 2 years ago

268bbecf

Fix forward AD add.Scalar to not set tangent as-is

pytorchmergebot committed 2 years ago

e213c6ad

Fix forward AD copy_ into same-sized tensor without fw grad

pytorchmergebot committed 2 years ago

a2d159e6

Remove python implementation for eigh meta

lezcano committed 2 years ago

648a6658

fx quant: refactor qconfig setting out of find_matches

vkuzo committed 2 years ago

7b4e92ac

Support dropout(nested tensor) (#79318)

YifanShenSZ committed 2 years ago

1b25aa67

Use nvfuser's suportted aten ops; Validate partitioner on torchbench

SherlockNoMad committed 2 years ago

75ec3bd1

[DOC] Missing line in serialization notes (#79454)

CompRhys committed 2 years ago

62ba548c

Newer Older