strisunshinewentingwang/pytorch

Pull Requests Commits

Change compiled_autograd tests to xfail instead of skip (#110348)

voznesenskym committed 1 year ago

06464a34

[DCP] Fix 'torch.cpu' has no attribute 'current_device' in checkpoint/optimizer.py (#110299)

wz337 committed 1 year ago

a5886487

[export] Add run_decomposition() function to ExportedProgram (#110236)

angelayi committed 1 year ago

13af952f

Add heuristic for when `evict_first` should be set (and some other minor things) (#108841)

Chillee committed 1 year ago

13681382

[onnx] support attn_mask fp16 type (#110306)

rui-ren committed 1 year ago

e4414716

[2/N] Enable clang-tidy to c10/test/*cpp (#110270)

cyyever committed 1 year ago

55905c4a

[2/N] Clean up CMake target linking (#109986)

cyyever committed 1 year ago

ef5ff790

[AOTInductor] Add non-default device test (#110024)

khabinov committed 1 year ago

669faab0

[vision hash update] update the pinned vision hash (#110344)

pytorchupdatebot committed 1 year ago

2bcae755

[AOTInductor] Add model runner to avoid using torch_extension (#110263)

khabinov committed 1 year ago

e8c0364f

[AOTInductor] ProxyExecutor supports Tuple of Tensor and List[Tensor] in returns (#110187)

SherlockNoMad committed 1 year ago

898656e9

[inductor][fbcode] Add -D C10_DISABLE_TENSORIMPL_EXTENSIBILITY to cpp_compile_command (#110122)

ColinPeppler committed 1 year ago

6bb448a2

Enable misc clang-tidy checks (#110283)

cyyever committed 1 year ago

d0ad848a

Skip launching kernels with zero grid in AOT Inductor (#110312)

aakhundov committed 1 year ago

2ead6c2f

[BE] Clean up trymerge code handling flaky failures (#110133)

huydhn committed 1 year ago

81a74457

[Dynamo] Add functional triton kernel wrapper (#110185)

oulgen committed 1 year ago

f7ba3e85

[CUDA][cudaMallocAsync] Improve `PYTORCH_CUDA_ALLOC_CONF` error message (#104891)

eqy committed 1 year ago

6b846584

[BE] [3/N] Use nested namespaces (#110314)

malfet committed 1 year ago

ad8aef0f

Small optimization to how we call flash-attention (#110324)

drisspg committed 1 year ago

8745d2d4

[Inductor] Enable the item() and nonzero() codegen test on CPU (#110262)

leslie-fang-intel committed 1 year ago

7eeb392e

Simplify the conditionals used for learning rate calculation for `ConstantLR` learning rate scheduler (#109785)

ancestor-mithril committed 1 year ago

e0be9ebc

[aotinductor] Fix a missing schema issue for repeat_interleave (#110105)

desertfire committed 1 year ago

993eea0e

[LTC] correct AdaptiveAvgPool3d channel dim index for shape inference (#109822)

davidgens-cerebras committed 1 year ago

ee0bff20

[BE] Use `std::make_unique` (#110298)

malfet committed 1 year ago

5a87477e

Revert "Make unbind() overrideable for NT subclass (#109122)"

pytorchmergebot committed 1 year ago

b083058e

MAINT: pytorchify torch._numpy tests: core/ and fft/ (#109815)

ev-br committed 1 year ago

1e95a1ae

[fuzzing result][fuzz_torch_jit_lite_interpreter] read-heap-use-after-free (size 8) in std::_Function_base::_M_empty() (#110289)

OctavianGuzu committed 1 year ago

9c7071b0

Revert "MAINT: pytorchify torch._numpy tests: core/ and fft/ (#109815)"

pytorchmergebot committed 1 year ago

f2d7faf4

Adding Backward Support for NestedTensors and FlashAttention (#97485)

drisspg committed 1 year ago

28d69d52

dynamic_shapes + retrace exported program (#110276)

avikchaudhuri committed 1 year ago

359c2a53

Newer Older