strisunshinewentingwang/pytorch

Pull Requests Commits

[SR] Factor logic that determines managed tensors out of MemoryPlanner (#68295)

Mike Iovine committed 2 years ago

cc46dc45

[Pytorch Edge] Make Tracer support xirp metal segmentation (#69328)

JacobSzwejbka committed 2 years ago

276cb8f5

Add OpInfos for combinations, cartesian_prod, sum_to_size, ldexp, as_strided (#68853)

saketh-are committed 2 years ago

a07ffe8d

Back out "Add efficient zero tensors" (#69327)

Mark Richardson committed 2 years ago

834bd313

fix for python 3.10 for gradient opinfo (#68113)

ankitaS11 committed 2 years ago

c572a603

Fix some usages of CUDA_VERSION (#69092)

mcarilli committed 2 years ago

572c3e31

[C10D] [Easy] Use pinned memory for HtoD copies in Reducer:: sync_bucket_indices (#69298)

Andrew Tulloch committed 2 years ago

dbc8d9c9

Add operator div (#68528)

Shirong Wu committed 2 years ago

e2c7bd08

Add support for C++ frontend wrapper on Linux (#69094)

malfet committed 2 years ago

bede18b0

THPStorage: Prefer intrusive_ptr over owning raw pointers (#69248)

peterbell10 committed 2 years ago

33c3c539

[fix] add range check for `k` kthvalue_cpu (#68863)

kshitij12345 committed 2 years ago

9f39a2de

.github: Fix ci workflows generation (#69329)

seemethere committed 2 years ago

cc85b689

ci: Migrate docs push to GHA (#69172)

seemethere committed 2 years ago

f786b03f

re-enable layer_norm in autodiff (#69007)

jjsjann123 committed 2 years ago

db5425bc

[testing] Ignore expected_regex in assertRaisesRegex for non-native device (#68723)

kshitij12345 committed 2 years ago

5b2586fe

Remove unused _convolution_nogroup op (#68829)

jbschlosser committed 2 years ago

36ba1b6b

[TensorExpr] Add lowerings for quantized ops: cat, mul, conv1d, relu. (#69055)

Mikhail Zolotukhin committed 2 years ago

791d5087

[TensorExpr] Add a pass to symbolize an input dimension. (#68857)

Mikhail Zolotukhin committed 2 years ago

83c4451f

[TensorExpr] TensorExprKernel: support custom-class constants. (#68856)

Mikhail Zolotukhin committed 2 years ago

1e9dcdd2

[TensorExpr] IR Eval: add more logging. (#68855)

Mikhail Zolotukhin committed 2 years ago

48d7d585

[TensorExpr] Un-const TEK::kernel_func_name to allow recompilation. (#68854)

Mikhail Zolotukhin committed 2 years ago

b6bcf5a0

Revert D32404517: [quant][embedding qat] Support Embedding QAT via FX API

malfet committed 2 years ago

a0367f89

Revert D32318435: [quant][embdding qat] Add FX support for QAT EmbeddingBag

malfet committed 2 years ago

ec4c7490

Forward fix merge conflict (#69319)

janeyx99 committed 2 years ago

8dafe6e1

Fix `ChainedScheduler.get_last_lr()` (#69112)

kurtamohler committed 2 years ago

52219b10

[pytorch][PR] bug fix for D32374003 (#69278)

Ming Zhu committed 2 years ago

db30696b

GHA: preserve downloaded JSONs as artifacts (#69258)

janeyx99 committed 2 years ago

915c26f5

Deprecate torch.triangular_solve (#63570)

lezcano committed 2 years ago

cafcf599

Expose MobileCode to python (#66592)

tugsbayasgalan committed 2 years ago

dde80168

Enabling CUDA 11.5 for binary builds, Adding windows workflows for CUDA 11.5 (#69262)

atalman committed 2 years ago

bb522c9d

Newer Older