strisunshinewentingwang/pytorch

Pull Requests Commits

optimize RNN on CPU (#22512)

xiaomengy committed 5 years ago

8bdda03a

jjsjann123 committed 5 years ago

3135298d

Improve hypothesis_utils.py for qtensor (#22693)

jerryzh168 committed 5 years ago

1682d38a

bddppq committed 5 years ago

3fabb9f1

add fill_diagonal function (#21892)

shihongzhi committed 5 years ago

45cf33a7

Add environment variables used in CONTRIBUTING example (#22736)

kianasun committed 5 years ago

89d6e880

enabled MIOpen depthwise convolutions (#22696)

lcskrishna committed 5 years ago

5147819f

Quantized Conv2d Module (#21323)

z-a-f committed 5 years ago

d21e476d

Mark Unpickler data ptr arg as const

mrshenli committed 5 years ago

ad634875

Revert D16183577: Delegate Python ~ (invert operator) to Tensor.bitwise_not().

colesbury committed 5 years ago

42402209

Revert D15998762: [jit] Give functions qualified names

Karl Ostmo committed 5 years ago

1ecc945a

Revert D15998758: [jit] refactor self to be a class again

Karl Ostmo committed 5 years ago

a1ca3240

Revert D16184799: [jit] make CompilationUnit::define return defined functions

Karl Ostmo committed 5 years ago

e6eb1730

izdeby committed 5 years ago

fffa7200

add a comment to native_functions explaining softmax interfaces (#22651)

nairbv committed 5 years ago

67c634d5

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

Newer Older