strisunshinewentingwang/pytorch

Pull Requests Commits

Revert "Remove device maps from TensorPipe for v1.7 release (#45353)" (#45762)

mrshenli committed 3 years ago

8cb72802

Make sure each warnings.warn only executes once inside TorchScript. (#45382)

gmagogsfm committed 3 years ago

d150d3e2

[caffe2] Optimize Dedup version of RowWiseSparseAdagrad fused op by WarpReduce (#45649)

xianjiec committed 3 years ago

73e9daa3

Torch Integration Test Formatting Changes (#45740)

MarcioPorto committed 3 years ago

c31066ac

Update backward definition for more operators and reenable tests in test_ops.py (#44444)

anjali411 committed 3 years ago

7d809f5d

[te] Get llvm codegen to compile with llvm9 and llvm-fb (#45726)

bertmaher committed 3 years ago

e3d2defd

Revert D24018160: [pytorch][PR] Test torch.svd using complex float and double numbers

malfet committed 3 years ago

5a47a212

Enable NamedTuple data type to work with DDP (#44220)

rohan-varma committed 3 years ago

f8c1ca5d

Fix cuDNN error message when it's Conv2d (#45729)

xwang233 committed 3 years ago

8619de84

type check for torch.quantization.observer (#45630)

Rong Rong committed 3 years ago

322855e3

Change signature for torch.poisson (#45656)

Ansley Ussery committed 3 years ago

db8b0762

Add function signature for pixel_shuffle (#45661)

Ansley Ussery committed 3 years ago

7726754e

adding sharding option to run_test.py (#45583)

janeyx99 committed 3 years ago

6acd7b68

[Docs] Adding Store API Docs (#45543)

osalpekar committed 3 years ago

3799ba83

Bump nightlies to 1.8.0 (#45696)

seemethere committed 3 years ago

a052597e

Use correct signatures for METH_NOARGS. (#45528)

pritamdamania committed 3 years ago

6e43f0db

Add string versions of argument funcs in jit Node (#45464)

spaugh committed 3 years ago

cdf93b03

Exposes SparseToDenseMask Caffe2 Operator (#45670)

MarcioPorto committed 3 years ago

b234acd4

Add a distributed package reviewer (#45744)

mingzhe09088 committed 3 years ago

ad31068f

Enable type check for torch.quantization.fake_quantize (#45701)

samestep committed 3 years ago

24187a0b

Test torch.svd using complex float and double numbers (#45572)

malfet committed 3 years ago

888f3c12

remove beta defaulting in smooth_l1_loss_backward. added to the bc whitelist (#45588)

bdhirsh committed 3 years ago

4d08930c

some documentation and style fixes to smooth_l1_loss (#45587)

bdhirsh committed 3 years ago

869b2ca0

make broadcasting explanation clearer in matmul doc: #22763 (#45699)

bdhirsh committed 3 years ago

c703602e

VariableKernel calls into scattered C++ api (#44158)

smessmer committed 3 years ago

82cc86b6

Add faithful C++ API (#44087)

smessmer committed 3 years ago

6e2eee2b

Use addmm directly for 1x1 convolution (#45557)

ngimel committed 3 years ago

9201c37d

[quant] PerChannelFloatQParams support for quint4x2 dtype (#45594)

supriyar committed 3 years ago

1a2d3b6a

[quant] creating quint4x2 dtype for quantized tensors (#44678)

supriyar committed 3 years ago

04526a49

Set the default bailout depth to 20 (#45710)

Krovatkin committed 3 years ago

a0d08b21

Newer Older