strisunshinewentingwang/pytorch

Pull Requests Commits

Fix the error in the note about `torch.device` documentation. (#16839)

JirenJin committed 5 years ago

a9f1d2e3

Register coalescer bug was fixed in ROCm 2.1 (#16923)

iotamudelta committed 5 years ago

19790b21

Alignas is now correctly handled on ROCm (#16920)

iotamudelta committed 5 years ago

d72c5d5a

Enable buildin bitonic sort (#16919)

iotamudelta committed 5 years ago

5089ee96

Change the default image size from 227 to 224 in resnet50 trainer (#16924)

bddppq committed 5 years ago

f169f398

enable unit tests working on ROCm 2.1 (#16871)

iotamudelta committed 5 years ago

23e1c55c

Add suggest add to __constants__ message on save fail

Elias Ellison committed 5 years ago

fc4f33b0

Make the exception raised from "numpy.dtype(numpy.void, (INT,))" less cryptic (#16809)

chandlerzuo committed 5 years ago

6737190b

Revert D13970381: [caffe2] Add visibility to registry class to fix ubsan error

bwasti committed 5 years ago

12bace14

Extend Net.RunAllOnGPU() to support RecurrentNetwork op (#15713)

malfet committed 5 years ago

0799a81c

delete critical section in TH*Tensor_addmm (#16889)

James Reed committed 5 years ago

48fe839d

Revert D13806753: [pytorch][PR] TensorIterator cuda launch configs update

bwasti committed 5 years ago

f83556bb

Allow sequential modules in module list (#16882)

Elias Ellison committed 5 years ago

cd2dca3c

Impl ExpandDims op and fallback to CPU if needed (#15264)

gujinghui committed 5 years ago

5ada54e0

Add visibility to registry class to fix ubsan error (#16792)

bwasti committed 5 years ago

54c981d9

Remove Legacy entry point. (#16721)

ezyang committed 5 years ago

b9b0be7a

Deduplicate instances caching allocator, so that we only have one instance. (#16720)

ezyang committed 5 years ago

b3fbd3ee

Delete duplicate copy of THCCachingAllocator (round two). (#16615)

ezyang committed 5 years ago

5c982622

Bump caffe2 docker images to 248 (#16863)

bddppq committed 5 years ago

f0329629

Also register op schema when no kernels are registered

smessmer committed 5 years ago

6ce147c0

Don't automatically handle context parameter (#16867)

smessmer committed 5 years ago

2c713032

Support onnxifi with partially shaped inferred net (#16877)

Yinghai Lu committed 5 years ago

fe5989d4

Robust determination of cudnn library and relevant conda packages. (#16859)

pearu committed 5 years ago

7ce33c58

Specialize LengthsRangeFill and SparseLengthsWeightedSum in bound shape inference (#16869)

Yinghai Lu committed 5 years ago

930ed00b

Activation histogram net observer with multiple histogram files as output (#16855)

csummersea committed 5 years ago

b5111918

Allow dicts in C++ frontend (#16846)

David Riazati committed 5 years ago

ee0e71be

Separate elementwise level2 math functions (#16753)

xiaomengy committed 5 years ago

2db847b3

Fix (#2) ppc64le build break on git status --porcelain check (#16852)

avmgithub committed 5 years ago

22477c6a

doc updates for TorchScript (#16866)

suo committed 5 years ago

96369506

Fix autodiff of nll_loss

asuhan committed 5 years ago

67bb7b29

Newer Older