strisunshinewentingwang/pytorch

Pull Requests Commits

add line numbers to jit_log.h

Krovatkin committed 5 years ago

0196e0ba

make CompilationUnit::define return defined functions

suo committed 5 years ago

c49a71f9

refactor self to be a class again (#22207)

suo committed 5 years ago

ee9c8a75

Give functions qualified names (#22206)

suo committed 5 years ago

c0674ceb

Move Quantization Models to common_quantization (#22706)

Lucas Kabela committed 5 years ago

86fc4171

Turn on USE_DIRECT_NVRTC in fbcode again. (#22685)

ezyang committed 5 years ago

ebafa2e1

register __getitem__ builtin

wanchaol committed 5 years ago

edeb4dbd

Fix a FIXME in test_nn (#22675)

ssnl committed 5 years ago

368dbb9a

Fix Trace inlining of graphs with optional inputs (#22686)

Elias Ellison committed 5 years ago

00df49c9

Add common_quantized test case utilities (#22694)

Lucas Kabela committed 5 years ago

3e3e6ee3

Refactor and improve randperm tests.

xuhdev committed 5 years ago

7750cae7

Swap detection order in randperm_out_cuda to avoid unnecessary conversion from float when the input is small.

xuhdev committed 5 years ago

32709af8

Support Half type in randperm.

xuhdev committed 5 years ago

0f7c3710

Delegate Python ~ (invert operator) to Tensor.bitwise_not().

xuhdev committed 5 years ago

9c4c9c3a

Add a bitwise NOT operator for integer and Boolean types (CUDA).

xuhdev committed 5 years ago

574e8086

Add a bitwise NOT operator for integer and Boolean types (CPU).

xuhdev committed 5 years ago

e2dc1fc7

Refactoring _wrap_outputs to remove python dependence.

mal committed 5 years ago

58e20638

fix dce over loops

suo committed 5 years ago

ec1b669d

skip import nccl and gloo_gpu in cpu machine (#22522)

xw285cornell committed 5 years ago

9b8d7717

fix print (#22689)

jerryzh168 committed 5 years ago

b984b0ab

Enable more passes in ProfilingGraphExecutor

Krovatkin committed 5 years ago

f81395b3

Added Bfloat16 tensor for cpu with very limited support (#21860)

izdeby committed 5 years ago

10c60b60

keep reuqires_grad unchanged after converting bn to syncbn (#22569)

tianzhi0549 committed 5 years ago

6eb3969a

Revert D16161144: [pytorch][PR] Add traces to LowerGradOf and SpecializeAutoGrad

ezyang committed 5 years ago

cbb0b816

Enabled BFloat16 storage (#21523)

izdeby committed 5 years ago

3a8d7463

Updating submodules

svcscm committed 5 years ago

932ec8aa

Intoducing bfloat16 type (#21522)

izdeby committed 5 years ago

e72b617e

add forward declaration in stateful dataset (#22562)

xzhu1900 committed 5 years ago

de5a481c

Enable C2 operators running with {cpu, gpu} * {forward, backward} (#22664)

mingzhe09088 committed 5 years ago

3cf5f22f

change c2 bench to use new tensor creation interface (#22663)

mingzhe09088 committed 5 years ago

95a5da17

Newer Older