strisunshinewentingwang/pytorch

Pull Requests Commits

Update NNPI backend to 0.5.1.8 (#4397)

omromano committed 4 years ago

59f923e8

[jit] Adding vectorized load/store support for JIT generated CUDA kernel (#36555)

FDecaYed committed 4 years ago

346215ca

Add RecordFunctionGuard (#36215)

Ilia Cherniavskii committed 4 years ago

3ae70cb8

Link NCCL lib to TORCH_PYTHON_LINK_LIBRARIES when USE_NCCL=1 (#36948)

Shen Li committed 4 years ago

a14a8376

Fix flaky test_barrier_timeout* tests for test_distributed. (#36963)

pritamdamania committed 4 years ago

32307efd

Add sync-point insertions and block/thread local memory allocations (#36563)

zheng-xq committed 4 years ago

5e504e83

Delete docker build job for pytorch-linux-bionic-clang9-thrift-llvmdev (#36930)

Ailing Zhang committed 4 years ago

0647f344

[quant][graph] Add quantizedmul_relu and quantized::mul_scalar_relu ops (#36844)

supriyar committed 4 years ago

c03d1494

[quant][graph] Support for quantized::mul and quantized::mul_scalar (#36818)

supriyar committed 4 years ago

ee2a9ac5

Add bundled input support to speed_benchmark_torch (#36765)

dreiss committed 4 years ago

1c15cb47

[quant] Minor refactor in fused conv names (#36883)

z-a-f committed 4 years ago

dbdd0f50

add absolute alias for abs (#36597)

jessebrizzi committed 4 years ago

28f439d4

hardsigmoid: add PyTorch wrapper for the QNNPACK path (#36699)

vkuzo committed 4 years ago

4d171c0e

hardsigmoid operator for QNNPACK, using LUTs (#36698)

vkuzo committed 4 years ago

1d720228

Canonicalize includes in aten, and add tests for it (#36301)

zasdfgbnm committed 4 years ago

97f2513c

Convert C casts to static casts (UnaryOpsKernel) (#36400)

kshitij12345 committed 4 years ago

47023148

[quant][graph] Update quantize_dynamic_script API to take sample model args (#36817)

supriyar committed 4 years ago

a2951a1e

Remove hacked twins from codegen (#36666)

smessmer committed 4 years ago

63e9d95c

[Distribution] Implement kl divergence for Cauchy distribution (#36477)

xidulu committed 4 years ago

4e365b9c

fix explicitly defaulted constexpr assignment operator fails to compile error for gcc 5.3.0 (#36561)

XiaobingSuper committed 4 years ago

4d2502a0

[quant][graphmode] tanh pattern and test (#36880)

z-a-f committed 4 years ago

e0e70589

[profiler] Allow record_function ctx manager to profile futures (#35055)

rohan-varma committed 4 years ago

752d3c28

Report error for lt, le, gt, ge in complex Vec256 (consistent with <, <=, >, >=) (#36646)

xuhdev committed 4 years ago

1e054bfb

ROCm: don't warn about CUDA compute capabilities (#35949)

t-vi committed 4 years ago

dc4d8881

add a fast path for EmbeddingBag calling FBGEMM (#36679)

xing-liu committed 4 years ago

2fa17ded

[rpc] Remove redundant call to createExceptionResponse (#36857)

rohan-varma committed 4 years ago

68f847c4

Add at::aten::hardsigmoid symbol (#36851)

JackCaoG committed 4 years ago

399f494d

ai-pep: match the qlinear benchmark to linear (#36674)

vkuzo committed 4 years ago

13391ceb

ai-pep: align qconv benchmark to conv (#36673)

vkuzo committed 4 years ago

25649684

Bmm sparse dense (#33430)

kurtamohler committed 4 years ago

c7cf4c1b

Newer Older