strisunshinewentingwang/pytorch

Pull Requests Commits

Enabled conv2d batch norm fusion for traced modules.

kimishpatel committed 4 years ago

8727f637

Add overloaded name to prim operators (#34280)

iseeyuan committed 4 years ago

d4f182d0

Removes unused THCTensor_(triu), THCTensor_(div) (#34712)

Mike Ruberry committed 4 years ago

c86d1361

solve conv3d backward get incorrect result problem (#34358)

XiaobingSuper committed 4 years ago

c258e473

Move min and max(reduce all) to Aten(CPU) (#33936)

XiaobingSuper committed 4 years ago

7848c229

Disallow sending CUDA tensors over RPC for current RPC agents. (#33604)

pritamdamania committed 4 years ago

f058c03b

CUDA Loops: move address computation into policy, make policy.load load all arguments (#33720)

zasdfgbnm committed 4 years ago

f404537c

Fix backward compatibility check test for schemas containing (#34782)

kimishpatel committed 4 years ago

528aabd3

[PyTorch BC] Clean up the BC whitelist (#34784)

houseroad committed 4 years ago

15c84c37

Remove unnecessary import (#34778)

tgrlnc committed 4 years ago

08bc3c6c

Export roi_align_gradient_op to c10 (#34776)

xiaomengy committed 4 years ago

1d81bd02

Fix missing header (#34762)

Yinghai Lu committed 4 years ago

373c80ee

Revert D20311699: [pytorch][PR] [C++ API] RNN / GRU / LSTM layer refactoring

Will Feng committed 4 years ago

6c555e15

Enable threading for XNNPACK ops. (#34547)

kimishpatel committed 4 years ago

84bd71db

Pass to remove prepacking ops. (#34319)

kimishpatel committed 4 years ago

4da55693

JIT pass to insert XNNPACK ops (#34048)

kimishpatel committed 4 years ago

7dd5da20

Integrate XNNPACK with custom class for packing weights. (#34047)

kimishpatel committed 4 years ago

4c30fc72

[C++ API] RNN / GRU / LSTM layer refactoring (#34322)

Will Feng committed 4 years ago

e23a9dc1

[reland][quant][graphmode] Add quantized conv2d-relu fusion pattern (#33279) (#34744)

jerryzh168 committed 4 years ago

5710374e

Move torchbind out of jit namespace (#34745)

James Reed committed 4 years ago

fb20621b

[quant][onnx] Support conversion of quantized sigmoid operator from pytorch to caffe2 (#34629)

supriyar committed 4 years ago

8a395882

[quant][onnx] Add support to convert max_pool2d quantized pytorch op to C2 (#33945)

supriyar committed 4 years ago

af289151

[C++ API] RNNCell / LSTMCell / GRUCell layers (#34400)

Will Feng committed 4 years ago

d041d078

Add the quantized batch_norm3d and also batch_norm3d fused with relu operators (#34702)

lly-zero-one committed 4 years ago

68758b2f

[C++ API] Link to module options doc for functional that has same options as module (#34752)

Will Feng committed 4 years ago

da11646d

.circleci: Remove CUDA 10.0, no longer needed (#34726)

seemethere committed 4 years ago

7dee36a0

invokeOperatorFromPython: support overloaded operator calling (#34671)

zdevito committed 4 years ago

52005b55

[JIT][mobile] Support built-in Function call in lite interpreter (#34676)

James Reed committed 4 years ago

ab76a820

[jit] small cleanups after script:: removal (#34677)

suo committed 4 years ago

af3a7e2b

[fix] use non-inplace for insert observer pass (#34190)

jerryzh168 committed 4 years ago

e7910aa9

Older