strisunshinewentingwang/pytorch

Pull Requests Commits

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

PyTorch Dockerfile should update submodules recursively.

xuhdev committed 5 years ago

4dbeb87e

conditionally defined var better error message (#20911)

Elias Ellison committed 5 years ago

0aeb9716

Add support for recursive compilation on Modules (#20708)

davidriazati committed 5 years ago

2f4824b2

Remove old custom op implementation (#21085)

smessmer committed 5 years ago

834d678e

Add aliasAnalysis to torch::RegisterOperators() (#21084)

smessmer committed 5 years ago

384d828e

c10::OperatorOptions (#21181)

smessmer committed 5 years ago

80556761

registration options should only be callable on rvalues (#21079)

smessmer committed 5 years ago

b91e0d14

Implement various AliasAnalysis operations directly on top of MemoryLocations. (#21203)

Owen Anderson committed 5 years ago

18179217

Cuda persistent softmax (#20827)

thorjohnsen committed 5 years ago

e098878d

Move legacy TH functions(sinh,cosh) to TensorIterator + Vec256 (#21115)

xta0 committed 5 years ago

052bab70

remove quantize_linear from Tensor method (#21196)

jerryzh168 committed 5 years ago

7f960a9c

remove dependency to caffe2::math and eigen (#21169)

jspark1105 committed 5 years ago

c185145d

improve test_docs_coverage error messages (#21029)

Brennan Vincent committed 5 years ago

8c927b20

Fix weak module cuda() _flat_weights bug (#21107)

davidriazati committed 5 years ago

e13b483f

introduce a new intrace to add op [C2 changes] (#21148)

mingzhe09088 committed 5 years ago

0223d374

introduce a new interface to add op [core changes] (#21147)

mingzhe09088 committed 5 years ago

31089b02

Revert D15454048: Move THCTensor_{normal, normal_means, normal_stddevs, normal_means_stddevs} to ATen

ezyang committed 5 years ago

012069ca

Newer Older