strisunshinewentingwang/pytorch

Pull Requests Commits

Delay loading the cuda library on Windows (#37811)

peterjc123 committed 4 years ago

0359a9b0

Move addmm broadcasting code from codegen layer to native layer. (#37613)

gchanan committed 4 years ago

91c1505e

Move addbmm broadcasting from the codegen layer to native layer. (#37603)

gchanan committed 4 years ago

6792c3ad

Revert D21406034: [pytorch][PR] [BE] Add @skipIfNoFBGEMM decorator

ezyang committed 4 years ago

b8d48d36

Fix weight quantization in RNNs (#35961)

raghuramank100 committed 4 years ago

34bf868e

Revert D21171334: [pytorch][PR] Change StorageImpl to track byte count rather than element count

ezyang committed 4 years ago

a2fc7f78

fix undef CUDA_VERSION warning (#37866)

EscapeZero committed 4 years ago

563bbeb8

reenable quantization test_qadd_scalar_relu test (#37423)

vkuzo committed 4 years ago

0cae7187

reenable quantized test_compare_tensor_scalar (#37422)

vkuzo committed 4 years ago

b61fda23

workaround nvcc host function bug (#37867)

EscapeZero committed 4 years ago

b57d82fc

[Tensorpipe Agent] Call Shutdown from Destructor and Join (#37839)

osalpekar committed 4 years ago

30a65f1a

Add zero_mask() for Vec256<BFloat16> (#37114)

xuhdev committed 4 years ago

5325606c

Make aten_tensor_iterator ASAN safe (#37869)

malfet committed 4 years ago

4c009c7f

[TensorExpr] Add a constructor accepting a name_hint to class Buf. (#36617)

Mikhail Zolotukhin committed 4 years ago

27fc2ab9

[TensorExpr] Add dtype to class Buf. (#36611)

Mikhail Zolotukhin committed 4 years ago

1c0bad25

[Testing] Add `--save-xml` option (#37840)

malfet committed 4 years ago

2c6aed0d

[Tensorpipe Agent] Adding Tensorpipe Codeowners (#37854)

osalpekar committed 4 years ago

a3639fa5

Change StorageImpl to track byte count rather than element count (#37776)

kurtamohler committed 4 years ago

3706803b

Fix `cdist` backward calculation for `p=2` (#37337)

Emilio Castillo committed 4 years ago

25ba802c

[BE] Add @skipIfNoFBGEMM decorator (#37810)

malfet committed 4 years ago

06e1b688

Remove unused capture in tensorpipe_agent.cpp (#37828)

Shen Li committed 4 years ago

65291fd4

[jit] fix trace checking reporting divergent names (#37842)

suo committed 4 years ago

bd220b33

[Caffe2] raise exceptions instead of str (#37744)

quangIO committed 4 years ago

9d7a79ac

[wip] Make quantization modules work with DataParallel (#37032)

vkuzo committed 4 years ago

b57c8b72

quant BN tests: remove qint32 (#37832)

vkuzo committed 4 years ago

25e6129c

add a cuda job for profiling tests (#37812)

Krovatkin committed 4 years ago

08304ccc

Tighten error checking in ConcreteModuleType (#37813)

suo committed 4 years ago

5b0244ee

Specify _th_ ops in CUDAUnaryOps macros so they are easier to find. (#37582)

gchanan committed 4 years ago

782b53b6

[quant][graphmode][refactor] rename SwapDequant and refactor code handling general ops (#37555)

jerryzh168 committed 4 years ago

9b3911c0

[TensorExpr] Add python bindings for TE fuser. (#37831)

Mikhail Zolotukhin committed 4 years ago

7fa968b1

Newer Older