strisunshinewentingwang/pytorch

Pull Requests Commits

Int8GivenTensorFill Operator Schema fix typo (#16204)

eyyub committed 5 years ago

e661dc27

Add support for fusion of half batch norm with float stats (#16735)

apaszke committed 5 years ago

8f6ee88a

Improve the Sparse matrix multiplication computational speed #16187 (#16905)

musikisomorphie committed 5 years ago

c282afff

Allow dataloader to accept a custom memory pinning function (#16743)

definitelynotmcarilli committed 5 years ago

07428746

Add abs for ByteTensor and CharTensor. (#16893)

hameerabbasi committed 5 years ago

73d7ecd1

Support named tuple return from operators on JIT (#16253)

zasdfgbnm committed 5 years ago

eae139e1

Enhance the documentation for torch.nn.DataParallel (#15993)

bombs-kim committed 5 years ago

9cb41e53

DOC: correct docstring for torch and torch.Tensor package (#16842)

ZhuBaohe committed 5 years ago

aae6b53c

find libnvToolsExt instead of using only hardcoded path (#16714)

t-vi committed 5 years ago

6a528007

Clean up autograd method tests

zasdfgbnm committed 5 years ago

8c9df48f

fixed LogSigmoid math string that wasn't rendering in documentation (#16900)

Travis Johnston committed 5 years ago

a1a330bd

ctc_loss error message bug fix. (#16917)

drkw1git committed 5 years ago

e0323a6a

Use non-Variable type for callsites that check type equality (#16325)

Will Feng committed 5 years ago

202eaa4e

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

Newer Older