strisunshinewentingwang/pytorch

Pull Requests Commits

move batchnorm and layernorm fusion to decompose (#20337)

wanchaol committed 5 years ago

871c9dcb

Quantized Conv2d operator (#20772)

dskhudia committed 5 years ago

cde611a6

add batch of string ops (#20826)

Elias Ellison committed 5 years ago

aebcd80a

make wildcards alias only each other (#20670)

suo committed 5 years ago

7aa3887f

Mark values entering containers as wildcards (#20556)

suo committed 5 years ago

90910fc6

Fix bug in exporting node with multiple outputs by scripting

BowenBao committed 5 years ago

28be521e

fix pow bug on overloads and clean up (#20824)

wanchaol committed 5 years ago

c2e3e79a

Allow both Variables and Tensors in c10 kernel interface (#20816)

smessmer committed 5 years ago

98928f4d

Add as_quantized_tensor (#20740)

jerryzh168 committed 5 years ago

9ea009fe

Change comparison ops result dtype to bool [Part1] (#20767)

izdeby committed 5 years ago

12bc81ae

Update references to minimum CUDA and cuDNN version (#20718)

colesbury committed 5 years ago

6ec3c122

CUDA implementation of fakequant (#20252)

jerryzh168 committed 5 years ago

05543153

Revert D15445092: Some minor fix to unblock the Bert model quantization

ezyang committed 5 years ago

fdb92399

fix add_histogram_raw (#20688)

lanpa committed 5 years ago

cfc98ae7

Exposing LengthsSum/Mean/Max in pytorch (#20802)

kittipatv committed 5 years ago

fd2aa93b

ONNX Export Scatter

lara-hdr committed 5 years ago

8d7a0257

Add ability to filter metric schema in LayerModelHelper (#20786)

shunting314 committed 5 years ago

fea4a56a

Automatic update of fbcode/onnx to cc2333a3f929caca7223b98699237f19388dd585 (#20763)

houseroad committed 5 years ago

810816a1

Fix optimizer type hint (#20648)

TheCodez committed 5 years ago

4e0d098a

When detecting numpy, assign relavant variables outside the try block (#20739)

xuhdev committed 5 years ago

795a1a6f

Revert D15248618: Split ATen/Parallel into interface and backend

ezyang committed 5 years ago

fd95947e

Some minor fix to unblock the Bert model quantization (#20787)

lly-zero-one committed 5 years ago

70ecddfd

Add quant-dequant nodes for bias. (#20045)

nishantpdce committed 5 years ago

a501e7d5

Add DictType to Metadata (#20770)

houseroad committed 5 years ago

c2d0e731

Use AT_INTERNAL_ASSERT in test_base (#20555)

suo committed 5 years ago

70eb315d

Rename FC to Linear in the test routine (#20716)

jianyuh committed 5 years ago

4a85e795

fbgemm precision argument (#20790)

jerryzh168 committed 5 years ago

77651615

Split ATen/Parallel into interface and backend (#20057)

Ilia Cherniavskii committed 5 years ago

c4a3b4d5

int_repr for different quantized types (#20656)

jerryzh168 committed 5 years ago

adbab828

Use SmallVector to allocate Compound operands inline. (#20762)

Owen Anderson committed 5 years ago

c1d6bcf3

Newer Older