strisunshinewentingwang/pytorch

Pull Requests Commits

Allow cublas an cudnn to be in different nvidia folders (#92122)

dannyjeck committed 1 year ago

a799acec

[Executorch][Quantization] Backend Config for functional embedding (#92700)

JacobSzwejbka committed 1 year ago

eb32bb2c

[SDPA] Integrating the main branch of flash_attn instead of cutlass (#91994)

drisspg committed 1 year ago

9613395e

where() function added as a Tensor method as well (#92849)

pierreHaslee committed 1 year ago

1c30844e

Revert #92688 and #92348 (aot autograd explicitly errors on double backward) (#92863)

soulitzer committed 1 year ago

fb980581

Remove unnecessary includes from `python_variable.cpp` (#92839)

kurtamohler committed 1 year ago

397b1a3d

Add missing moves to torch autograd (#92772)

Skylion007 committed 1 year ago

8c8cd953

ci: Increase timeout for linux binary builds (#92859)

seemethere committed 1 year ago

2a8669c5

Add Meta backend into tensor type strings (#92697)

GlebKazantaev committed 1 year ago

402c6d42

[PT-D][Checkpoint]rename init() (#92829)

wz337 committed 1 year ago

dd4b46e0

Update XLA pin (#92806)

lezcano committed 1 year ago

7560660b

[lint] clang-format register_prim_ops_fulljit.cpp (#92150)

IvanKobzarev committed 1 year ago

57fe3340

Revert "Run all of the timm models shards in the periodic (#92743)"

pytorchmergebot committed 1 year ago

2cf03bbb

Remove deprecated torch.symeig (#70988)

IvanYashchuk committed 1 year ago

d70ed681

[caffe2] Remove OperatorBase::newstyle_outputs_ (#67093)

peterbell10 committed 1 year ago

dd251112

Splitting #91254 into two PRs (#92748)

Fabio Rocha committed 1 year ago

e137dcc2

[PT-D][Checkpoint]Resolve issue #89501: Rename _nested_tensor.py to (#92705)

wz337 committed 1 year ago

f7e1f3e8

Add CUDA 11.8 CI workflows (#92137)

ptrblck committed 1 year ago

9bfd1357

Create pt2_bug_report.yml (#92773)

ngimel committed 1 year ago

f3338857

Move ASAN and ONNX to Python 3.9 and 3.8 (#92712)

malfet committed 1 year ago

3643d5de

[ONNX] Support ListConstruct in quantized_args (#92009)

titaiwangms committed 1 year ago

4e9539e0

[SDPA] Update SDPA API and make function Public (#92189)

drisspg committed 1 year ago

df14650f

Allow direct Tensor constructor to return preexisting PyObject (#92754)

ezyang committed 1 year ago

1237cf6b

Added vectorized horizontal flip path for channels last for NcHW (#91806)

vfdev-5 committed 1 year ago

e994e783

Simplify retains grad hook implementation (#92604)

soulitzer committed 1 year ago

a112814a

[Docker] Factor GHCR push into its own step (#92832)

malfet committed 1 year ago

71b10512

sparse_sparse_matmul: simplify backward (#91712)

nikitaved committed 1 year ago

9f381c9b

[BE]: remove old dataclasses install from CI (#92763)

Skylion007 committed 1 year ago

36ba2ce5

A few usability improvements for the dynamo benchmarks. (#92713)

Fabio Rocha committed 1 year ago

a43b55e1

Fix typo under docs directory (#92762)

kiszk committed 1 year ago

d40a4540

Newer Older