strisunshinewentingwang/pytorch

Pull Requests Commits

Enable CUDA 11 jobs for Windows nightly builds (#44086)

peterjc123 committed 4 years ago

538d3bd3

[quant] conv_transpose2d_prepack/conv_transpose2d_unpack (#40351)

z-a-f committed 4 years ago

69e38828

[android][jni] Support Tensor MemoryFormat in java wrappers (#40785)

IvanKobzarev committed 4 years ago

c40e3f9f

[NNC] make inlining immediate (take 2) and fix bugs (#43885)

nickgg committed 4 years ago

70aecd2a

[TVM] Support Fused8BitRowwiseQuantizedToFloat op (#44098)

Hao Lu committed 4 years ago

bc4a00c1

[TensorExpr] Fuser: rely on input types when checking whether a device is supported. (#44139)

Mikhail Zolotukhin committed 4 years ago

3105d8a9

fx qat: respect device affinity (#44115)

vkuzo committed 4 years ago

71510c60

[JIT] Add mypy type annotations for JIT (#43862)

Meghan Lele committed 4 years ago

7816d537

[jit] Better match behavior of loaded ScriptModules vs. freshly created ones (#43298)

suo committed 4 years ago

9dd8670d

[jit] fix segfault in attribute lookup on loaded ScriptModules (#43284)

suo committed 4 years ago

74f18476

[tensorexpr] Alias analysis tests (#44110)

bwasti committed 4 years ago

e64879e1

[JIT] Fuser match on schemas not node kind (#44083)

Elias Ellison committed 4 years ago

6868bf95

[pytorch] Make mobile find_method return an optional (#43965)

ann-ss committed 4 years ago

9b3c72d4

Enable function calls in TEFuser and SpecializeAutogradZero (#43866)

Krovatkin committed 4 years ago

f91bdbea

[quant] Prep for conv_transpose packing (#39714)

z-a-f committed 4 years ago

e05fa2f5

[caffe2] fix clang build

Igor Sugak committed 4 years ago

352a32e7

Enable Enum pickling/unpickling. (#43188)

gmagogsfm committed 4 years ago

f3da9e3b

Benchmarks: add scripts for FastRNNs results comparison. (#44134)

Mikhail Zolotukhin committed 4 years ago

d0421ff1

Polish DDP join API docstrings (#43973)

rohan-varma committed 4 years ago

3806c939

Enable typechecks for torch.nn.modules.[activation|upsampling] (#44093)

malfet committed 4 years ago

442684cb

Fix replaceAtenConvolution for BC. (#44036)

kimishpatel committed 4 years ago

a153f694

Fix transposed conv2d rewrite pattern to account for convolution api (#44035)

kimishpatel committed 4 years ago

ba65cce2

Test TE fuser unary ops and fix sigmoid(half) (#44094)

bertmaher committed 4 years ago

55ff9aa1

Update rocm-3.5.1 build job to rocm-3.7 (#44123)

malfet committed 4 years ago

bfa1fa52

For CriterionTests, have check_gradgrad actually only affect gradgrad checks. (#44060)

gchanan committed 4 years ago

49215d7f

Mark bucketize as not subject to autograd (#44102)

t-vi committed 4 years ago

42f98979

add tanh + quantize unit test (#44076)

Hector Yuen committed 4 years ago

91b0d186

[JIT] Improve error message for unsupported Optional types (#44054)

Meghan Lele committed 4 years ago

de672e87

[TensorExpr] Benchmarks: set number of profiling runs to 2 for PE. (#44112)

Mikhail Zolotukhin committed 4 years ago

d11603de

[pytorch][bot] update mobile op deps (#44100)

ljk53 committed 4 years ago

b10c527a

Newer Older