-
Notifications
You must be signed in to change notification settings - Fork 5.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add momentum operator #4571
Merged
Merged
Add momentum operator #4571
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,89 @@ | ||
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. | ||
|
||
Licensed under the Apache License, Version 2.0 (the "License"); | ||
you may not use this file except in compliance with the License. | ||
You may obtain a copy of the License at | ||
|
||
http://www.apache.org/licenses/LICENSE-2.0 | ||
|
||
Unless required by applicable law or agreed to in writing, software | ||
distributed under the License is distributed on an "AS IS" BASIS, | ||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
See the License for the specific language governing permissions and | ||
limitations under the License. */ | ||
|
||
#include "paddle/operators/momentum_op.h" | ||
|
||
namespace paddle { | ||
namespace operators { | ||
|
||
class MomentumOp : public framework::OperatorWithKernel { | ||
public: | ||
using framework::OperatorWithKernel::OperatorWithKernel; | ||
|
||
protected: | ||
void InferShape(framework::InferShapeContextBase *ctx) const override { | ||
PADDLE_ENFORCE(ctx->HasInput("Param"), | ||
"Input(param) of Momentum should not be null."); | ||
PADDLE_ENFORCE(ctx->HasInput("Grad"), | ||
"Input(grad) of Momentum should not be null."); | ||
PADDLE_ENFORCE(ctx->HasInput("Velocity"), | ||
"Input(velocity) of Momentum should not be null."); | ||
PADDLE_ENFORCE(ctx->HasInput("LearningRate"), | ||
"Input(LearningRate) of Momentum should not be null."); | ||
|
||
PADDLE_ENFORCE(ctx->HasOutput("ParamOut"), | ||
"Output(ParamOut) of Momentum should not be null."); | ||
PADDLE_ENFORCE(ctx->HasOutput("VelocityOut"), | ||
"Output(VelocityOut) of Momentum should not be null."); | ||
|
||
auto param_dim = ctx->GetInputDim("Param"); | ||
PADDLE_ENFORCE_EQ( | ||
param_dim, ctx->GetInputDim("Grad"), | ||
"Param and Grad input of MomentumOp should have the same dimension."); | ||
PADDLE_ENFORCE_EQ( | ||
param_dim, ctx->GetInputDim("Velocity"), | ||
"Param and Velocity of MomentumOp should have the same dimension."); | ||
PADDLE_ENFORCE_EQ(framework::product(ctx->GetInputDim("LearningRate")), 1, | ||
"Learning_rate should be a scalar"); | ||
|
||
ctx->SetOutputDim("ParamOut", param_dim); | ||
ctx->SetOutputDim("VelocityOut", param_dim); | ||
} | ||
}; | ||
|
||
class MomentumOpMaker : public framework::OpProtoAndCheckerMaker { | ||
public: | ||
MomentumOpMaker(framework::OpProto *proto, | ||
framework::OpAttrChecker *op_checker) | ||
: OpProtoAndCheckerMaker(proto, op_checker) { | ||
AddInput("Param", "Input parameter"); | ||
AddInput("Grad", "Input gradient"); | ||
AddInput("Velocity", "Input velocity"); | ||
AddInput("LearningRate", "Input learning rate"); | ||
|
||
AddOutput("ParamOut", "Output parameter"); | ||
AddOutput("VelocityOut", "Output velocity"); | ||
|
||
AddAttr<float>("mu", "Momentum coefficient"); | ||
AddComment(R"DOC( | ||
|
||
Momentum Algorithm (momentum). | ||
|
||
velocity_out = mu * velocity - learning_rate * grad | ||
param_out = param + velocity_out | ||
|
||
Ref: Sutskever, Ilya, et al. "On the importance of initialization | ||
and momentum in deep learning." ICML 2013; | ||
http://jmlr.org/proceedings/papers/v28/sutskever13.pdf | ||
|
||
)DOC"); | ||
} | ||
}; | ||
} // namespace operators | ||
} // namespace paddle | ||
|
||
namespace ops = paddle::operators; | ||
REGISTER_OP_WITHOUT_GRADIENT(momentum, ops::MomentumOp, ops::MomentumOpMaker); | ||
REGISTER_OP_CPU_KERNEL( | ||
momentum, ops::MomentumOpKernel<paddle::platform::CPUPlace, float>); |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,20 @@ | ||
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. | ||
|
||
Licensed under the Apache License, Version 2.0 (the "License"); | ||
you may not use this file except in compliance with the License. | ||
You may obtain a copy of the License at | ||
|
||
http://www.apache.org/licenses/LICENSE-2.0 | ||
|
||
Unless required by applicable law or agreed to in writing, software | ||
distributed under the License is distributed on an "AS IS" BASIS, | ||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
See the License for the specific language governing permissions and | ||
limitations under the License. */ | ||
|
||
#define EIGEN_USE_GPU | ||
#include "paddle/operators/momentum_op.h" | ||
|
||
namespace ops = paddle::operators; | ||
REGISTER_OP_GPU_KERNEL( | ||
momentum, ops::MomentumOpKernel<paddle::platform::GPUPlace, float>); |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,53 @@ | ||
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve. | ||
|
||
Licensed under the Apache License, Version 2.0 (the "License"); | ||
you may not use this file except in compliance with the License. | ||
You may obtain a copy of the License at | ||
|
||
http://www.apache.org/licenses/LICENSE-2.0 | ||
|
||
Unless required by applicable law or agreed to in writing, software | ||
distributed under the License is distributed on an "AS IS" BASIS, | ||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
See the License for the specific language governing permissions and | ||
limitations under the License. */ | ||
|
||
#pragma once | ||
#include "paddle/framework/eigen.h" | ||
#include "paddle/framework/op_registry.h" | ||
|
||
namespace paddle { | ||
namespace operators { | ||
|
||
using Tensor = framework::Tensor; | ||
template <typename T, int MajorType = Eigen::RowMajor, | ||
typename IndexType = Eigen::DenseIndex> | ||
using EigenVector = framework::EigenVector<T, MajorType, IndexType>; | ||
|
||
template <typename Place, typename T> | ||
class MomentumOpKernel : public framework::OpKernel<T> { | ||
public: | ||
void Compute(const framework::ExecutionContext& ctx) const override { | ||
auto param_out = ctx.Output<Tensor>("ParamOut"); | ||
auto velocity_out = ctx.Output<Tensor>("VelocityOut"); | ||
|
||
param_out->mutable_data<T>(ctx.GetPlace()); | ||
velocity_out->mutable_data<T>(ctx.GetPlace()); | ||
|
||
float mu = ctx.Attr<float>("mu"); | ||
|
||
auto p = EigenVector<T>::Flatten(*ctx.Input<Tensor>("Param")); | ||
auto g = EigenVector<T>::Flatten(*ctx.Input<Tensor>("Grad")); | ||
auto v = EigenVector<T>::Flatten(*ctx.Input<Tensor>("Velocity")); | ||
float lr = ctx.Input<Tensor>("LearningRate")->data<float>()[0]; | ||
auto p_out = EigenVector<T>::Flatten(*param_out); | ||
auto v_out = EigenVector<T>::Flatten(*velocity_out); | ||
auto place = ctx.GetEigenDevice<Place>(); | ||
|
||
v_out.device(place) = mu * v - lr * g; | ||
p_out.device(place) = p + v_out; | ||
} | ||
}; | ||
|
||
} // namespace operators | ||
} // namespace paddle |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,35 @@ | ||
import unittest | ||
import numpy as np | ||
from op_test import OpTest | ||
|
||
|
||
class TestMomentumOp(OpTest): | ||
def setUp(self): | ||
self.op_type = "momentum" | ||
|
||
param = np.random.random((123, 321)).astype("float32") | ||
grad = np.random.random((123, 321)).astype("float32") | ||
velocity = np.zeros((123, 321)).astype("float32") | ||
learning_rate = np.array([0.001]).astype("float32") | ||
mu = 0.0001 | ||
|
||
self.inputs = { | ||
'Param': param, | ||
'Grad': grad, | ||
'Velocity': velocity, | ||
'LearningRate': learning_rate | ||
} | ||
|
||
self.attrs = {'mu': mu} | ||
|
||
velocity_out = mu * velocity - learning_rate * grad | ||
param_out = param + velocity_out | ||
|
||
self.outputs = {'ParamOut': param_out, 'VelocityOut': velocity_out} | ||
|
||
def test_check_output(self): | ||
self.check_output() | ||
|
||
|
||
if __name__ == "__main__": | ||
unittest.main() |
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That might be not good for GPU. If the
LearningRate
is in GPU memory, we cannot getfloat
directly.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay.. Thanks.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Fixed as per #4598