strisunshinewentingwang/pytorch

Pull Requests Commits

replace tl.math with tl. libdevice

alanwaketan committed 1 year ago

452a7463

Skip testing distributed backend if the backend (UCC, NCCL, Gloo) is not available (#98576)

huydhn committed 1 year ago

5bcbb9bc

[dynamo 3.11] enable dynamo unittests in 3.11 (#98104)

williamwen42 committed 1 year ago

117da58b

[caffe2] Micro-optimizations in BlobGetMutableTensor (#98103)

swolchok committed 1 year ago

457afe48

Assert that there are not duplicate sources for distinct arguments (#98738)

ezyang committed 1 year ago

02cff647

[better_engineering][multiplatform] Replace host_info() check with separate cmd and cmd_exe commands for protos (#98426)

kkolur76 committed 1 year ago

b663f7e8

[torch.library] Add ability to create library fragments (#98439)

zou3519 committed 1 year ago

d5120ff1

Fix test_python_dispatch under debug mode (#98609)

zou3519 committed 1 year ago

618ea6fa

[autograd_function_db] Add NumpyTake as OpInfo (#98438)

zou3519 committed 1 year ago

01b2c456

Skip failing test_torchinductor_codegen_dynamic_shapes tests on CPU (#98621)

huydhn committed 1 year ago

c139df40

Make all Source subclasses frozen. (#98737)

ezyang committed 1 year ago

9abae6ae

[CUDA12] set_device change (#94864)

Aidyn-A committed 1 year ago

69eef5a4

Avoid passing buffers to optimizers during spmd rematerialization (#98714)

mrshenli committed 1 year ago

3fcc5ff0

fix backward bug for custom device (#98586)

heidongxianhua committed 1 year ago

a3701b67

feat(add method is_private_use1() in class Device) (#98123)

ykddd committed 1 year ago

537c3461

Convert logging f-strings to use % format, part two (#98700)

ezyang committed 1 year ago

b09722f5

Convert logging f-strings to use % format (#98697)

ezyang committed 1 year ago

9a8f71f2

[xla hash update] update the pinned xla hash (#98195)

pytorchmergebot committed 1 year ago

ad88afcf

[aot] fix disable amp for runtime wrapper (#97864)

Valentine233 committed 1 year ago

95621b3c

Turn off cudagraph trees (#98709)

eellison committed 1 year ago

96fb64a1

[vision hash update] update the pinned vision hash (#98654)

pytorchmergebot committed 1 year ago

fdfd370c

use float as accumulate type for reduce Ops: min, max, minmax on CPU (#96079)

mingfeima committed 1 year ago

58424446

Fused attention patterns (#97741)

jansel committed 1 year ago

8fee4669

Improve dynamo support for autograd.Function (#98158)

jansel committed 1 year ago

f4858fa8

[JIT] Allow `tuple` and `list` generics (#98703)

malfet committed 1 year ago

7e0c26d4

distinguish mutability of TensorImpl::data() (#97776)

mikey dagitses committed 1 year ago

2400cb1d

Removed hip call hipDeviceSynchronize (#97209)

BLOrange-AMD committed 1 year ago

6b9a1cf8

[primTorch] add ref for `cumprod` (#98670)

nkaretnikov committed 1 year ago

ff825de4

make TensorImpl::data_ptr_impl() non-const and have mutable in the name (#97744)

mikey dagitses committed 1 year ago

9d363616

Support LayerNorm without weight or bias parameters (#98687)

mrshenli committed 1 year ago

54b16848

Older