strisunshinewentingwang/pytorch

Pull Requests Commits

Move RecordFunction into ATen (#37548)

Ilia Cherniavskii committed 4 years ago

2d708cef

Make RecordFunction callbacks thread local and modernize interface (#37491)

Ilia Cherniavskii committed 4 years ago

c24c5f96

Move resize / zero logic for _thnn_conv_depthwise2d from codegen to native code. (#37957)

gchanan committed 4 years ago

dc251908

Move _thnn_conv2d resize and zero code from codegen to native code. (#37956)

gchanan committed 4 years ago

ed4e7cec

Fixed gradcheck for complex (#37836)

anjali411 committed 4 years ago

99349393

Remove unpacked int8 blob after constructing the packed blob to save memory (#37973)

csummersea committed 4 years ago

8a8b7a16

[Tensorpipe Agent] Implementing getMetrics with currently available metrics (#37980)

osalpekar committed 4 years ago

f0f58736

[Tensorpipe Agent] Network Data Profiling (#37852)

osalpekar committed 4 years ago

5d21a9cf

[Tensorpipe Agent] Implement Global Interpreter Lock Wait Time Metric (#37851)

osalpekar committed 4 years ago

25359f73

[Tensorpipe Agent] Base Structs for Tracking RPC Metrics (#37850)

osalpekar committed 4 years ago

b452fef5

[TensorExpr] Allow to enable/disable fallback mechanism thru an envvar PYTORCH_TENSOREXPR_FALLBACK. (#37971)

Mikhail Zolotukhin committed 4 years ago

a44824c9

[TensorExpr] Move controlling knob out of the TE fuser pass. (#37970)

Mikhail Zolotukhin committed 4 years ago

067f08c1

Remove overly strict assertion for type demotion of scalars. (#38001)

resistor committed 4 years ago

3066d3ac

[quant] Release qnnpack original weights for conv/linear (#37595)

supriyar committed 4 years ago

7bf9d983

Make speed_benchmark_torch report latency in us (#37953)

kimishpatel committed 4 years ago

dd64d26d

Message Delay fix for test_check_failed_messages (#37978)

osalpekar committed 4 years ago

85fccba2

Update miniconda repository, be specific about cudatoolkit (#37186)

seemethere committed 4 years ago

305444a0

[BE] Add @skipIfNoFBGEMM decorator (Reland) (#37894)

malfet committed 4 years ago

2b41b9bc

[CUDA_FUSER] Expand operation support for cuda fuser (#37849)

jjsjann123 committed 4 years ago

1667aa64

[TensorPipe] Update submodule (#38013)

lw committed 4 years ago

ffed9dca

Move resize logic for bmm from codegen to native code. (#37955)

gchanan committed 4 years ago

b2cc9928

Acquire GIL when constructing/destructing ConcretePyObjectHolder (#37870)

mrshenli committed 4 years ago

ee1ddcef

Add support for non-persistent buffers. (#37191)

sharvil committed 4 years ago

594b33ea

Add --check-untyped-defs to mypy.ini and test suite (#37594)

rgommers committed 4 years ago

46ed3349

Migrate CUDA where, tril, triu to c10::complex (#37896)

zasdfgbnm committed 4 years ago

30fc58cf

[ONNX] Support clamp_min and clamp_max (#37872)

BowenBao committed 4 years ago

7be9796c

[TensorPipe] Use the new multi-payload message API (#37919)

lw committed 4 years ago

bc09478a

[TensorPipe] Allow passing args to agent options constructor (#37918)

lw committed 4 years ago

978ad162

remove deprecation warning on get_contiguous_memory_format (#37963)

suo committed 4 years ago

4e93844a

Fix splitWithTail to insert the tail immediately after the outer loop. (#37941)

resistor committed 4 years ago

65260d48

Newer Older