Skip to content

Enable cast operator for a NG bridge#17294

Merged
tensor-tang merged 1 commit intoPaddlePaddle:developfrom
mozga-intel:mozga-intel/cast_operator
May 29, 2019
Merged

Enable cast operator for a NG bridge#17294
tensor-tang merged 1 commit intoPaddlePaddle:developfrom
mozga-intel:mozga-intel/cast_operator

Conversation

@mozga-intel
Copy link
Contributor

@mozga-intel mozga-intel commented May 8, 2019

Enable the cast operator for the nGraph bridge. BERT model uses it.

@mozga-intel mozga-intel requested review from luotao1 and tensor-tang May 8, 2019 19:34
@mozga-intel mozga-intel force-pushed the mozga-intel/cast_operator branch from 8ba8fc2 to 062c857 Compare May 9, 2019 08:09
@mozga-intel mozga-intel added this to the v1.5 for Intel milestone May 14, 2019
@mozga-intel mozga-intel force-pushed the mozga-intel/cast_operator branch 2 times, most recently from 48dda4e to ef9c9f1 Compare May 20, 2019 21:45
@mozga-intel
Copy link
Contributor Author

@luotao1 Could you tell me what kind of test is it: PR_CI_Document_Preview (Compile) — TeamCity build failed

@mozga-intel
Copy link
Contributor Author

@luotao1 I see that the following test is not required. Can I skip it?

@luotao1
Copy link
Contributor

luotao1 commented May 21, 2019

You can skip it.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

empty line.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

@mozga-intel mozga-intel force-pushed the mozga-intel/cast_operator branch 2 times, most recently from 6e14682 to 815ac76 Compare May 23, 2019 02:02
@mozga-intel mozga-intel force-pushed the mozga-intel/cast_operator branch 2 times, most recently from 9c49548 to 8a42606 Compare May 24, 2019 21:40
Copy link
Contributor

@baojun-nervana baojun-nervana left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

from __future__ import print_function

import unittest
from paddle.fluid.tests.unittests.test_cast_op import TestCastOp1
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why just only import the first one?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

testcase2 and 3 involve fp16 which is not supported.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK

Copy link
Contributor

@tensor-tang tensor-tang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

from __future__ import print_function

import unittest
from paddle.fluid.tests.unittests.test_cast_op import TestCastOp1
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK

@tensor-tang tensor-tang merged commit ccf9e23 into PaddlePaddle:develop May 29, 2019
@baojun-nervana baojun-nervana deleted the mozga-intel/cast_operator branch May 29, 2019 05:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants