OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-07 12:21:27 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity
19,178 Commits 8,304 Branches 150 Tags 1.5 GiB
b3147bc674
Commit Graph

7 Commits

Author SHA1 Message Date
BowenBao
b3147bc674 PyTorch export to ONNX Opset 7 and 8 - Cont (#22421) 
Summary:
This is an extension to the original PR https://github.com/pytorch/pytorch/pull/21765

1. Increase the coverage of different opsets support, comments, and blacklisting.
2. Adding backend tests for both caffe2 and onnxruntime on opset 7 and opset 8.
3. Reusing onnx model tests in caffe2 for onnxruntime.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/22421

Reviewed By: zrphercule

Differential Revision: D16225518

Pulled By: houseroad

fbshipit-source-id: 01ae3eed85111a83a0124e9e95512b80109d6aee
 2019-07-12 14:52:48 -07:00
Lara
42c6ea5faa ONNX Export Topk with Dynamic k (+ add test cases) 
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/21104

Differential Revision: D16061592

Pulled By: houseroad

fbshipit-source-id: 855b310a138fdde9c25869ffe9f127189dc2eaf5
 2019-07-05 23:46:36 -07:00
Lara Haidar
7ca7edc307 ONNX Export LayerNorm 
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/22265

Reviewed By: zrphercule

Differential Revision: D16076268

Pulled By: houseroad

fbshipit-source-id: 29b4ecab2fa0dc7250c9d1ad6924903181a66ab2
 2019-07-02 09:37:07 -07:00
Chris Seymour
d8de69d621 Adds symbolic op for logsumexp 
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/22306

Differential Revision: D16046027

Pulled By: houseroad

fbshipit-source-id: 7319fd58321220941250c5b8eff024914798e392
 2019-06-29 00:09:06 -07:00
Lara
45c6fa0007 Refactor Tests for Multiple ONNX Opsets (#20036) 
Summary:
Refactor tests for https://github.com/pytorch/pytorch/pull/19294.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20036

Reviewed By: zrphercule

Differential Revision: D16016593

Pulled By: houseroad

fbshipit-source-id: eaae324e347679acf3d0ac1c14be03919f54496e
 2019-06-26 17:06:57 -07:00
Lara
7b1ffba3bf ArgumentStash for Scalar arguments (#21931) 
Summary:
Scalars are being traced as constants.
This PR is to fix this issue.

The ONNX Graph for Test_Full_op() before and after this change:

def Test_Full_op():
  class Test_Full(nn.Module):
    def forward(self, x):
      return torch.full((3, 4), x, dtype=torch.long)
  model = Test_Full()
  x = torch.tensor(12)
  output = model(x)

Before this change:
graph(%input1 : Long()):
%output1 : Float(3, 4) = onnx::Constant[value=<Tensor>]
return (%output1)

After this change:
graph(%input1 : Long()):
%1 : int[] = onnx::Constant[value= 3 4 [ Variable[CPULongType]{2} ]]
%2 : Tensor = onnx::ConstantOfShape[value={0}]
%output1 : Float(3, 4) = onnx::Add(%2, %input1)
return (%output1)

Similar PR : https://github.com/pytorch/pytorch/pull/12939
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21931

Reviewed By: zrphercule

Differential Revision: D15950066

Pulled By: houseroad

fbshipit-source-id: 3470665d88fa34faa600940ef16b069a06002cd5
 2019-06-25 15:22:08 -07:00
Lu Fang
c1744a6c39 Add ONNX py3 CI cases (#21715) 
Summary:
So far, we only have py2 ci for onnx. I think py3 support is important. And we have the plan to add onnxruntime backend tests, which only supports py3.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21715

Reviewed By: bddppq

Differential Revision: D15796885

Pulled By: houseroad

fbshipit-source-id: 8554dbb75d13c57b67ca054446a13a016983326c
 2019-06-14 10:20:14 -07:00