strisunshinewentingwang/pytorch

Pull Requests Commits

Port register_prim_ops_c10.cpp to new registration API (#37834)

mjavanmard committed 4 years ago

53e7d49a

[JIT] rename enable_profiling_mode to enable_profiling_mode_for_profiling_tests (#37825)

Elias Ellison committed 4 years ago

0e3a05ec

Migrate check_convert to c10::complex (#37875)

zasdfgbnm committed 4 years ago

436cd2c0

modify `select_equals_backward` to propage only to a single value (#36316)

nikitaved committed 4 years ago

84342476

[JIT]Support adv indexing using list. (#37848)

Ailing Zhang committed 4 years ago

dd618216

Revert D21409626: [quant][tests] Enable tests to run on all qengine backends

Mike Ruberry committed 4 years ago

f2148de9

[docs] Fix broken links in `contribution_guide.rst` and `governance.rst` (#37820)

ShawnZhong committed 4 years ago

ec7fd0ca

Remove requantization scale constraint. (#37683)

kimishpatel committed 4 years ago

e729db48

Fix lint (#37922)

ezyang committed 4 years ago

6f06df81

[RESUBMIT] Kill broadcasting from the codegen layer. (#37907)

gchanan committed 4 years ago

122d8215

Change DeprecationWarning to UserWarning in `torch.cuda` (#32142)

jma127 committed 4 years ago

88c447bf

[quant][tests] Enable tests to run on all qengine backends (#37843)

supriyar committed 4 years ago

f78d02ed

Add Aten as dep to fakelowp and cpuinfo path to its include path (#37909)

Tal Cherckez committed 4 years ago

2f61b045

Fix some amount of support for Bool in tensorexpr. (#37914)

resistor committed 4 years ago

75c201ac

Support c10::optional<Tensor> in custom C++ autograd function. (#37700)

pritamdamania committed 4 years ago

cdc56d0b

Reduction should not coalesce_dimensions when splitting for 32bit indexing (#37788)

zasdfgbnm committed 4 years ago

b57b596f

Updating submodules

svcscm committed 4 years ago

222fdd42

Revert D21393512: [quant][graphmode] Support a new category of ops in graph mode quantization

jerryzh168 committed 4 years ago

ad2305e5

Back out "Revert D21171334: [pytorch][PR] Change StorageImpl to track byte count rather than element count" (#37893)

ezyang committed 4 years ago

fe888067

[TensorExpr] Fix the shape info check in the TE fuser pass. (#37882)

Mikhail Zolotukhin committed 4 years ago

8c91b782

[ROCm] Enable bfloat16 for ops in BERT model (#37634)

rohithkrn committed 4 years ago

e3934dfa

Enable ahead of time compilation for HIPExtensions using ninja (#37800)

ashishfarmer committed 4 years ago

402f635b

[quant] ConvPackedParams with TorchBind (#35923)

jerryzh168 committed 4 years ago

70f375be

Devirtualize device init calls in factory op wrappers (#37815)

Basil Hosmer committed 4 years ago

32b09f7a

[Caffe2] Increase timing threshold to 50 ms on Windows (#37892)

malfet committed 4 years ago

9f060d38

[quant][graphmode] Support a new category of ops in graph mode quantization (#37515)

jerryzh168 committed 4 years ago

5eacc9cb

Stop defining static data in Vec256 (#37767)

ezyang committed 4 years ago

480bd0ad

fix msan in vec_reduce_all (#37853)

EscapeZero committed 4 years ago

96b512be

Revert D21310335: reenable quantization test_qadd_scalar_relu test

Ailing Zhang committed 4 years ago

e3d1c4ea

disable clang-tidy modernize-trailing-return (#37888)

Krovatkin committed 4 years ago

92f750b5

Newer Older