strisunshinewentingwang/pytorch

Pull Requests Commits

Python Dispatcher integration with C++ dispatcher (#84826)

voznesenskym committed 2 years ago

35f6a691

[quant][docs] Add example for the error message for fixed qparam ops (#84666)

jerryzh168 committed 2 years ago

44c30c5d

Remove enable_recursive_torch_dispatch (#84945)

ezyang committed 2 years ago

55ca297d

Removes unnecessary namespace of functions used only in einsum (#84955)

janeyx99 committed 2 years ago

922560b8

[vision hash update] update the pinned vision hash (#84975)

pytorchmergebot committed 2 years ago

d26e9cd9

[torchdynamo hash update] update the pinned torchdynamo hash (#84912)

pytorchmergebot committed 2 years ago

b28d82cb

Fix `expectedFailureMeta` to avoid skipping tests (#84875)

kurtamohler committed 2 years ago

d00cabae

Avoid nested CommTensor wrapping (#84963)

mrshenli committed 2 years ago

8cbbd3a2

Revert "Don't detach when making views; force caller to detach (#84893)"

pytorchmergebot committed 2 years ago

8ca057eb

[ROCm] re-enable tensorexpr and test_openmp (#81367)

arindamroy-eng committed 2 years ago

a185dc2e

Updated library level maintainers for torchtext (#84950)

Nayef211 committed 2 years ago

cb9ef466

[CMake] Add functorch target (#83464)

malfet committed 2 years ago

d05a1133

`ReflectionPad` supports `BFloat16` (#84949)

crcrpar committed 2 years ago

26b59862

[Profiler] Make `LibKinetoClient::stop()` directly call `ProfilerStateBase::pop` (#83965)

Taylor Robie committed 2 years ago

fdd36654

Don't detach when making views; force caller to detach (#84893)

ezyang committed 2 years ago

3bb8d6a9

Create Cache for Fusion Reuse in NVFuser in Python Frontend for Primtorch (#83267)

kevinstephano committed 2 years ago

ec916bf6

Adds mruberry as superuser (#84869)

mruberry committed 2 years ago

59bb5c93

disable onednn gelu for empty input (#84926)

XiaobingSuper committed 2 years ago

c61e8954

Updating cudnn_frontend to 0.7.1 (#84943)

atalman committed 2 years ago

25d91e0a

Revert "[reland] Call jit decomposition in VariableType to increase forward AD coverage (#84151) (#84675)"

pytorchmergebot committed 2 years ago

36d79143

Add __all__ for a few distributed modules plus a little typing (reland) (#84872)

Rodrigo Kumpera committed 2 years ago

38192f63

[functorch] test - vmapjvpvjp (#83375)

kshitij12345 committed 2 years ago

53c71e21

[ROCm] Remove gfx900 from base docker build and Pytorch build scripts (#80015)

jithunnair-amd committed 2 years ago

b4a881af

Revert D34636039: Multisect successfully blamed D34636039 for test or build failures (#84942)

jianyuh committed 2 years ago

0e8c5cf8

Return device count using nvml (#84879)

malfet committed 2 years ago

81da50a9

Memoize `torch.cuda.device_count` (#84878)

malfet committed 2 years ago

94f20c35

[Nested Tensor] Create differentiable nt to tensor view functions (#83371)

drisspg committed 2 years ago

bda8a572

Fix intermittent link errors in NCCL build (#84245)

peterbell10 committed 2 years ago

fa86874b

Don't use reentrant dispatch for composite compliance (#84909)

ezyang committed 2 years ago

74d0c647

autograd: fix non-deterministic output in codegen comments (#84695)

krallin committed 2 years ago

b4799736

Newer Older