strisunshinewentingwang/pytorch

Pull Requests Commits

Add gelu gradient for pytorch (#21237)

xiaomengy committed 5 years ago

31c79b71

Add gelu activation in pytorch (#20665)

xiaomengy committed 5 years ago

93ae040f

Revert D15577342: [pytorch][PR] Fix Python DataParallel RNN in no_grad mode

Karl Ostmo committed 5 years ago

aac424a6

Fixes a bug in the test (#21146)

Zafar Takhirov committed 5 years ago

360e6d1b

Exclude file:line from graphs used for fuser kernel cache (#21252)

James Reed committed 5 years ago

62ae348d

Save the weight shape info the first time we have chance to extract it (#21233)

Yinghai Lu committed 5 years ago

7c40576c

Revert D15548138: Export feature length information for onnxifi operator

Jiang Wu committed 5 years ago

0efc527d

Fix Python DataParallel RNN in no_grad mode (#21197)

mrshenli committed 5 years ago

51ebbe97

Fix typo in test_dataloader

ssnl committed 5 years ago

f051fbd4

compare scalar device with common device (#21236)

Natalia Gimelshein committed 5 years ago

d168a853

Fix wrong type hints for Tensor.is_cuda, is_leaf (#21192)

sublee committed 5 years ago

41b17e24

Fix `sccache not being used on Windows` (#21248)

peterjc123 committed 5 years ago

be7fc406

Add file-line info for jit.load and string frontend (#21217)

James Reed committed 5 years ago

619261d7

Lazily build error strings in schema matching using replay. (#21241)

Owen Anderson committed 5 years ago

b663eec1

fix contribution and governance links (#21243)

sovvo committed 5 years ago

5bc7c1f8

Export feature length information for onnxifi operator (#21110)

ChunliF committed 5 years ago

85786bea

add PT maxpool and avgpool ops to the benchmark suite (#21200)

mingzhe09088 committed 5 years ago

516ea33f

add PT conv and convtranspose ops to the benchmark suite (#21199)

mingzhe09088 committed 5 years ago

dceea734

add PT linear op to the benchmark suite (#21204)

mingzhe09088 committed 5 years ago

2d75d313

add PT batchnorm op to the benchmark suite (#21201)

mingzhe09088 committed 5 years ago

00b3e692

migrate matmul operator to the new interface (#21198)

mingzhe09088 committed 5 years ago

ed1078bd

avoid multiple writes to files on export (#21186)

suo committed 5 years ago

c8dc707f

Register gradient op with engine (#21205)

bddppq committed 5 years ago

4c19421f

Add file:line:graph to graph printout (#21180)

James Reed committed 5 years ago

daa1e2de

use _sparse_coo_tensor_unsafe in coalesce for speedup (#21214)

Aapo Kyrola committed 5 years ago

678dc44d

Reuse common options between ONNXIFI and TVM transformations (#21163)

Yinghai Lu committed 5 years ago

9e5f1db6

schema_matching.cpp: mark internal functions as static. (#21140)

Mikhail Zolotukhin committed 5 years ago

b12a5f61

migrate intraop benchmarks to the new interface (#21202)

mingzhe09088 committed 5 years ago

668dbcc4

migrate add operator to the new interface (#21152)

mingzhe09088 committed 5 years ago

c62d4762

remaining use of t.quantize_linear (#21219)

jerryzh168 committed 5 years ago

fd19d06d

Newer Older