Skip to content
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

[Pten] Gru lstm migration #39729

Merged
merged 7 commits into from
Mar 2, 2022

Conversation

iclementine
Copy link

@iclementine iclementine commented Feb 18, 2022

PR types

Others

PR changes

Others

Describe

  1. Move LSTM compute related files
paddle/fluid/operators/math/lstm_compute.h   =>   paddle/phi/kernels/funcs/lstm_compute.h
paddle/fluid/operators/math/lstm_compute.cc  =>  paddle/phi/kernels/funcs/lstm_compute.cc
paddle/fluid/operators/math/lstm_compute.cu  =>  paddle/phi/kernels/funcs/lstm_compute.cu
paddle/fluid/operators/math/detail/lstm_kernel.h   =>   paddle/phi/kernels/funcs/detail/lstm_kernel.h
paddle/fluid/operators/math/detail/lstm_cpu_kernel.h  =>  paddle/phi/kernels/funcs/detail/lstm_cpu_kernel.h
paddle/fluid/operators/math/detail/lstm_gpu_kernel.h  =>  paddle/phi/kernels/funcs/detail/lstm_gpu_kernel.h
  1. Move GRU compute related files;
paddle/fluid/operators/math/gru_compute.h   =>   paddle/phi/kernels/funcs/gru_compute.h
paddle/fluid/operators/math/gru_compute.cc  =>  paddle/phi/kernels/funcs/gru_compute.cc
paddle/fluid/operators/math/gru_compute.cu  =>  paddle/phi/kernels/funcs/gru_compute.cu
paddle/fluid/operators/math/detail/gru_kernel.h   =>   paddle/phi/kernels/funcs/detail/gru_kernel.h
paddle/fluid/operators/math/detail/gru_cpu_kernel.h  =>  paddle/phi/kernels/funcs/detail/gru_cpu_kernel.h
paddle/fluid/operators/math/detail/gru_gpu_kernel.h  =>  paddle/phi/kernels/funcs/detail/gru_gpu_kernel.h
  1. Move a small Activation functions system only used in LSTM and GRU;
paddle/fluid/operators/math/detail/activation_functions.h => paddle/phi/kernels/funcs/detail/activation_functions.h
paddle/fluid/operators/math/detail/avx_functions.cc => paddle/phi/kernels/funcs/detail/avx_functions.cc
paddle/fluid/operators/math/detail/avx_mathfun.h => paddle/phi/kernels/funcs/detail/avx_mathfun.h
  1. Move sequence2batch related files;
paddle/fluid/operators/math/sequence2batch.h  => paddle/phi/kernels/funcs/sequence2batch.h
paddle/fluid/operators/math/sequence2batch.cc  => paddle/phi/kernels/funcs/sequence2batch.cc
paddle/fluid/operators/math/sequence2batch.cu  => paddle/phi/kernels/funcs/sequence2batch.cu
  1. update all references to symbols within them.

@paddle-bot-old
Copy link

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

Copy link
Contributor

@chenwhql chenwhql left a comment

Choose a reason for hiding this comment

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

LGTM,一些命名空间后续PR可以追加修改为phi下面的

LstmMetaValue<T> value, LstmMetaGrad<T> grad,
int frame_size, int batch_size, T cell_clip,
ActivationType active_node, ActivationType active_gate,
void gpu_lstm_backward(const paddle::platform::DeviceContext& context,
Copy link
Contributor

Choose a reason for hiding this comment

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

paddle::platform::DeviceContext -> phi::DeviceContext

@@ -237,21 +298,37 @@ void gpu_lstm_backward(const platform::DeviceContext& context, Op op,
}

auto stream =
reinterpret_cast<const platform::CUDADeviceContext&>(context).stream();
reinterpret_cast<const paddle::platform::CUDADeviceContext&>(context)
Copy link
Contributor

Choose a reason for hiding this comment

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

paddle::platform::CUDADeviceContext -> phi::GPUContext

@@ -59,32 +61,37 @@ class LoDTensor2BatchFunctor {

public:
void operator()(const DeviceContext& context,
const framework::LoDTensor& lod_tensor,
framework::LoDTensor* batch, bool is_cal_batch_lod,
const paddle::framework::LoDTensor& lod_tensor,
Copy link
Contributor

Choose a reason for hiding this comment

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

paddle::framework::LoDTensor -> DenseTensor

Copy link
Contributor

@XiaoguangHu01 XiaoguangHu01 left a comment

Choose a reason for hiding this comment

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

LGTM

@iclementine iclementine merged commit e4dba69 into PaddlePaddle:develop Mar 2, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants