strisunshinewentingwang/pytorch

Pull Requests Commits

[JIT] Include custom_class.h in torch/script.h

James Reed committed 4 years ago

34ccfba4

[JIT] Serialize attributes and types in ClassType serialization

James Reed committed 4 years ago

06c19263

[JIT] Support for registering C++ lambdas as methods on custom C++ class

James Reed committed 4 years ago

1719da13

.circleci: Add workflows for Python 3.8 (#31948)

seemethere committed 4 years ago

da390914

consider FAIL_GUARD while counting indices for GUARDs (#32672)

Krovatkin committed 4 years ago

0dc38be4

Python binding to export bytecode format for lite interpreter (#32621)

Martin Yuan committed 4 years ago

c64dec19

Kill some more unused code in function_wrapper.py

gchanan committed 4 years ago

e24ce0e5

Fix spelling errors

hjung0 committed 4 years ago

9a2691f2

[jit] fix segfault on missing getstate (#32642)

suo committed 4 years ago

63170431

div_kernel: throw when dividing by integer zero (#32629)

Baranowski committed 4 years ago

8e416151

Fix flaky test_nccl_timeout. (#32653)

pritamdamania committed 4 years ago

b3848c56

[JIT] Fix classes as attributes in recursive scripting

James Reed committed 4 years ago

d68592a4

Use the C++ current RpcAgent pointer to eliminate the unnecessary argument passing from Python world (#32635)

xush6528 committed 4 years ago

b9f764b1

Clean up mvlgamma doc (including a weird way to link to reference) (#32667)

xuhdev committed 4 years ago

666e5430

Back out "Make autogen functions correct for multiple outputs and views" (#32681)

albanD committed 4 years ago

db8ce7ea

Make C++ RpcAgent::currentRPCAgent_ the source of truth of current RPC Agent (#32633)

xush6528 committed 4 years ago

5c8535d5

Updating submodules

svcscm committed 4 years ago

1217c9b3

Make _wait_all_workers() support being called for multiple times (#32624)

xush6528 committed 4 years ago

16959153

[vulkan][caffe2] Add logging for descriptor extensions, fp16 storage

IvanKobzarev committed 4 years ago

39987de9

[quantization] FP16 dynamic quantized Linear

James Reed committed 4 years ago

812b1ad8

Updating submodules

svcscm committed 4 years ago

389b9c18

Revert "Fix iterator for ncclCommWatchdog. (#32571)" (#32649)

ezyang committed 4 years ago

57519bd8

Kill THIntegerTensor, THDenseTensor, THDenseIndexTensor. (#32599)

gchanan committed 4 years ago

897b6908

Adding native qconcat

z-a-f committed 4 years ago

f6c46df8

Revert D19562258: [pytorch][PR] Fixes moving after weight norm application

ezyang committed 4 years ago

f0917dce

Back out "Use simd version for fp16 conversions" (#32640)

Yinghai Lu committed 4 years ago

64323ae1

Fixes moving after weight norm application (#32563)

Mike Ruberry committed 4 years ago

e36cbb8f

[ROCm] Adjust elementwise_kernel settings on ROCm (#32609)

iotamudelta committed 4 years ago

5ac2593d

0-dim batch size input for interpolate. (#32400)

v0dro committed 4 years ago

ca9dc670

verify input sizes for instance norm and group norm (#29082)

v0dro committed 4 years ago

602394e9

Newer Older