strisunshinewentingwang/pytorch

Pull Requests Commits

Move hardtanh activation to Aten(CPU, CUDA) (#30152)

XiaobingSuper committed 4 years ago

82c3f486

Revert D18117070: Migrate max and min (binary) from TH to ATen.

ezyang committed 4 years ago

6e38d503

we should have a config-based way to skip flaky tests (#29944)

zdevito committed 4 years ago

e5bd7a79

Fix error checking of CUDA multi_margin_loss. (#30825)

gchanan committed 4 years ago

0974dcc2

Revert "Default to not build Caffe2 operators on Windows. (#29061)" (#30740)

ezyang committed 4 years ago

2ced81f2

Vectorize smooth L1 loss backward function on CPU. (#30046)

xuhdev committed 4 years ago

f874230d

Fix `os.register_at_fork` not defined on Windows (#30809)

peterjc123 committed 4 years ago

6486bdfb

Add ATen/native/ headers to torch target (#30835)

zrphercule committed 4 years ago

c564d794

Add docs for how we expose declarations in at:: to torch:: (#30760)

Will Feng committed 4 years ago

244b0bd1

style fixes to code analyzer (#30808)

ljk53 committed 4 years ago

be55874f

Wrap warning handler in a function to avoid siof (#30800)

Qi Zhou committed 4 years ago

9617d07b

add torch_cpu to the static library list in TorchConfig.cmake.in (#30769)

ljk53 committed 4 years ago

bf1b4b6f

Deprecate tensor.type() (#30281)

ngoldbaum committed 4 years ago

f5318155

reenable cuda_kernel_loop_overflow_large test (#30797)

ngimel committed 4 years ago

2171f910

Migrate max and min (binary) from TH to ATen. (#27185)

xuhdev committed 4 years ago

1578a286

Fully deprecate variadic inputs of checkpoint_sequential (#25985)

sublee committed 4 years ago

fa251cfd

Turn off scalar_checks for SpatialDepthwiseConvolution and SpatialConvolutionMM. (#30789)

gchanan committed 4 years ago

26077729

Move scalar_check from codegen to code in MultiLabelMarginCriterion. (#30770)

gchanan committed 4 years ago

f12332eb

Fix scalar check of MultiLabelMarginLoss. (#30768)

gchanan committed 4 years ago

50625798

Fix a CUDA memory leak in MultiLabelMarginCriterion error checking. (#30767)

gchanan committed 4 years ago

473a0448

Turn off scalar_check for is_target for MultiLabelMarginCriterion, which is handled correctly in code. (#30766)

gchanan committed 4 years ago

ba1a9871

Support 0-d tensors in CUDA MultiLabelMarginCriterion. (#30765)

gchanan committed 4 years ago

35a69978

Provide full path for buck hipification (#30746)

Serhat Yilmaz committed 4 years ago

c4e9748b

CUDA-strided-complex Binary and Unary Op support (#30295)

dylanbespalko committed 4 years ago

f2a2fec4

Clean up non-C++14 code (#28443)

smessmer committed 4 years ago

139aa519

fix AvgPool2d for 2^31-1 sized inputs, and get test_cuda_kernel_loop_… (#30771)

soumith committed 4 years ago

a939b52d

Make `InsertQuantDeQuantHelper` global (#30550)

jerryzh168 committed 4 years ago

1d20c32b

Supporting making submodules unique (#30037)

jerryzh168 committed 4 years ago

c4c2e233

Stop producing op_version_set version numbers.

zdevito committed 4 years ago

7a2889b0

Invoke more passes in `insertObservers` (#30473)

jerryzh168 committed 4 years ago

3c1bb21c

Newer Older