wehub-resource-sync
364 lines
15 KiB
Plaintext
364 lines
15 KiB
Plaintext
// Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
|
|
//
|
|
// 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/phi/kernels/cudnn_lstm_grad_kernel.h"
|
|
|
|
#include "paddle/phi/core/kernel_registry.h"
|
|
#include "paddle/phi/kernels/gpu/cudnn_lstm_utils.h"
|
|
|
|
namespace phi {
|
|
|
|
template <typename T, typename Context>
|
|
void CudnnLSTMGradKernel(
|
|
const Context &dev_ctx,
|
|
const DenseTensor &x,
|
|
const DenseTensor &init_h,
|
|
const DenseTensor &init_c,
|
|
const optional<std::vector<const DenseTensor *>> &weight_list,
|
|
const optional<DenseTensor> &sequence_length,
|
|
const DenseTensor &out,
|
|
const DenseTensor &reserve,
|
|
const DenseTensor &state_out,
|
|
const DenseTensor &out_grad,
|
|
const DenseTensor &last_h_grad,
|
|
const DenseTensor &last_c_grad,
|
|
float dropout_prob,
|
|
bool is_bidirec,
|
|
int hidden_size,
|
|
int num_layers,
|
|
bool is_test,
|
|
int seed,
|
|
DenseTensor *x_grad,
|
|
DenseTensor *init_h_grad,
|
|
DenseTensor *init_c_grad,
|
|
std::vector<DenseTensor *> weight_grad_list) {
|
|
auto input_dims = x.dims();
|
|
auto init_h_dims = init_h.dims();
|
|
auto init_c_dims = init_c.dims();
|
|
|
|
auto *init_h_data = init_h.data<T>();
|
|
auto *init_c_data = init_c.data<T>();
|
|
auto *out_data = out.data<T>();
|
|
auto *out_grad_data = out_grad.data<T>();
|
|
auto *last_h_grad_data = last_h_grad.data<T>();
|
|
auto *last_c_grad_data = last_c_grad.data<T>();
|
|
|
|
auto running_weight_list = *weight_list.get_ptr();
|
|
int weight_numel = size_sum(running_weight_list);
|
|
bool continuous =
|
|
is_continuous<T, std::vector<const DenseTensor *>>(running_weight_list);
|
|
|
|
auto handle = dev_ctx.cudnn_handle();
|
|
auto place = dev_ctx.GetPlace();
|
|
auto stream = dev_ctx.stream();
|
|
DenseTensor weight_whole;
|
|
T *weight_data = nullptr;
|
|
|
|
if (!continuous) {
|
|
weight_whole.Resize({weight_numel});
|
|
dev_ctx.template Alloc<T>(&weight_whole);
|
|
weight_to_tensor<T>(place, stream, running_weight_list, &weight_whole);
|
|
weight_data = weight_whole.data<T>();
|
|
} else {
|
|
weight_data = const_cast<T *>(running_weight_list[0]->data<T>());
|
|
}
|
|
|
|
DenseTensor weight_grad;
|
|
funcs::SetConstant<GPUContext, T> zero;
|
|
weight_grad.Resize({weight_numel});
|
|
dev_ctx.template Alloc<T>(&weight_grad);
|
|
zero(dev_ctx, &weight_grad, static_cast<T>(0.0));
|
|
T *weight_grad_data = weight_grad.data<T>();
|
|
|
|
int offset = 0;
|
|
for (size_t i = 0; i < weight_grad_list.size(); ++i) {
|
|
size_t len = weight_grad_list[i]->numel();
|
|
auto dim = weight_grad_list[i]->dims();
|
|
weight_grad_list[i]
|
|
->ShareDataWith(weight_grad.Slice(static_cast<int64_t>(offset),
|
|
static_cast<int64_t>(offset + len)))
|
|
.Resize(dim);
|
|
offset += len;
|
|
}
|
|
|
|
x_grad->Resize(input_dims);
|
|
dev_ctx.template Alloc<T>(x_grad);
|
|
auto *in_grad_data = x_grad->data<T>();
|
|
|
|
if (init_h_grad) {
|
|
init_h_grad->Resize(init_h_dims);
|
|
dev_ctx.template Alloc<T>(init_h_grad);
|
|
}
|
|
auto *init_h_grad_data = init_h_grad ? init_h_grad->data<T>() : nullptr;
|
|
|
|
if (init_c_grad) {
|
|
init_c_grad->Resize(init_c_dims);
|
|
dev_ctx.template Alloc<T>(init_c_grad);
|
|
}
|
|
auto *init_c_grad_data = init_c_grad ? init_c_grad->data<T>() : nullptr;
|
|
|
|
auto running_seq_length = sequence_length.get_ptr();
|
|
bool has_seq_length = running_seq_length != nullptr;
|
|
std::vector<int> SequenceLength;
|
|
if (has_seq_length) {
|
|
SequenceLength = GetVectorFromTensor<int>(running_seq_length);
|
|
}
|
|
|
|
int seq_length = input_dims[0];
|
|
int64_t batch_size = x.dims()[1];
|
|
int64_t input_size = x.dims()[2];
|
|
// TODO(large-tensor): cudnn rnn dims not support int64
|
|
PADDLE_ENFORCE_LE_INT_MAX(batch_size, "batch_size");
|
|
PADDLE_ENFORCE_LE_INT_MAX(input_size, "input_size");
|
|
int batch_size_int = static_cast<int>(batch_size);
|
|
int input_size_int = static_cast<int>(input_size);
|
|
|
|
size_t workspace_size;
|
|
size_t reserve_size;
|
|
|
|
ScopedRNNBase rnn(seq_length,
|
|
batch_size_int,
|
|
input_size_int,
|
|
hidden_size,
|
|
num_layers,
|
|
dropout_prob,
|
|
seed,
|
|
weight_numel,
|
|
true,
|
|
is_bidirec);
|
|
|
|
rnn.Create<T>(handle,
|
|
dev_ctx.GetPlace(),
|
|
SequenceLength,
|
|
&workspace_size,
|
|
&reserve_size,
|
|
const_cast<DenseTensor *>(&state_out));
|
|
|
|
DenseTensor workspace_data_;
|
|
workspace_data_.Resize({static_cast<int64_t>(workspace_size)});
|
|
dev_ctx.template Alloc<uint8_t>(&workspace_data_);
|
|
const uint8_t *reserve_data = reserve.data<uint8_t>();
|
|
|
|
#if CUDNN_VERSION >= 90000
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardData_v8(handle,
|
|
rnn.rnn_desc(),
|
|
nullptr,
|
|
rnn.y_seq_desc(),
|
|
out_data,
|
|
out_grad_data,
|
|
rnn.x_seq_desc(),
|
|
in_grad_data,
|
|
rnn.init_h_desc(),
|
|
init_h_data,
|
|
last_h_grad_data,
|
|
init_h_grad_data,
|
|
rnn.init_c_desc(),
|
|
init_c_data,
|
|
last_c_grad_data,
|
|
init_c_grad_data,
|
|
rnn.weights_size(),
|
|
weight_data,
|
|
workspace_size,
|
|
workspace_data_.data<uint8_t>(),
|
|
reserve_size,
|
|
const_cast<uint8_t *>(reserve_data)));
|
|
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardWeights_v8(handle,
|
|
rnn.rnn_desc(),
|
|
CUDNN_WGRAD_MODE_ADD,
|
|
nullptr,
|
|
rnn.x_seq_desc(),
|
|
x.data<T>(),
|
|
rnn.init_h_desc(),
|
|
init_h.data<T>(),
|
|
rnn.y_seq_desc(),
|
|
out.data<T>(),
|
|
rnn.weights_size(),
|
|
weight_grad_data,
|
|
workspace_size,
|
|
workspace_data_.data<uint8_t>(),
|
|
reserve_size,
|
|
const_cast<uint8_t *>(reserve_data)));
|
|
#else
|
|
|
|
if (!has_seq_length) {
|
|
// This interface is used when the input/output is unpadded.
|
|
#ifdef PADDLE_WITH_HIP
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::miopenRNNBackwardData(handle,
|
|
rnn.rnn_desc(),
|
|
seq_length,
|
|
rnn.y_descs(),
|
|
out_data,
|
|
rnn.y_descs(),
|
|
out_grad_data,
|
|
rnn.last_h_desc(),
|
|
last_h_grad_data,
|
|
rnn.last_c_desc(),
|
|
last_c_grad_data,
|
|
rnn.weight_desc(),
|
|
weight_data,
|
|
rnn.init_h_desc(),
|
|
init_h_data,
|
|
rnn.init_c_desc(),
|
|
init_c_data,
|
|
rnn.x_descs(),
|
|
in_grad_data,
|
|
rnn.init_h_desc(),
|
|
init_h_grad_data,
|
|
rnn.init_c_desc(),
|
|
init_c_grad_data,
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::miopenRNNBackwardWeights(handle,
|
|
rnn.rnn_desc(),
|
|
seq_length,
|
|
rnn.x_descs(),
|
|
x.data<T>(),
|
|
rnn.init_h_desc(),
|
|
init_h.data<T>(),
|
|
rnn.y_descs(),
|
|
out.data<T>(),
|
|
rnn.weight_desc(),
|
|
weight_grad_data,
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
#else
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardData(handle,
|
|
rnn.rnn_desc(),
|
|
seq_length,
|
|
rnn.y_descs(),
|
|
out_data,
|
|
rnn.y_descs(),
|
|
out_grad_data,
|
|
rnn.last_h_desc(),
|
|
last_h_grad_data,
|
|
rnn.last_c_desc(),
|
|
last_c_grad_data,
|
|
rnn.weight_desc(),
|
|
weight_data,
|
|
rnn.init_h_desc(),
|
|
init_h_data,
|
|
rnn.init_c_desc(),
|
|
init_c_data,
|
|
rnn.x_descs(),
|
|
in_grad_data,
|
|
rnn.init_h_desc(),
|
|
init_h_grad_data,
|
|
rnn.init_c_desc(),
|
|
init_c_grad_data,
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardWeights(handle,
|
|
rnn.rnn_desc(),
|
|
seq_length,
|
|
rnn.x_descs(),
|
|
x.data<T>(),
|
|
rnn.init_h_desc(),
|
|
init_h.data<T>(),
|
|
rnn.y_descs(),
|
|
out.data<T>(),
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
rnn.weight_desc(),
|
|
weight_grad_data,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
#endif
|
|
} else {
|
|
#if !defined(PADDLE_WITH_HIP) && CUDNN_VERSION >= 7201
|
|
// for train
|
|
// This interface is used when the input/output is padded.
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardDataEx(handle,
|
|
rnn.rnn_desc(),
|
|
rnn.y_seq_desc(),
|
|
out_data,
|
|
rnn.y_seq_desc(),
|
|
out_grad_data,
|
|
nullptr,
|
|
nullptr,
|
|
rnn.last_h_desc(),
|
|
last_h_grad_data,
|
|
rnn.last_c_desc(),
|
|
last_c_grad_data,
|
|
rnn.weight_desc(),
|
|
weight_data,
|
|
rnn.init_h_desc(),
|
|
init_h_data,
|
|
rnn.init_c_desc(),
|
|
init_c_data,
|
|
rnn.x_seq_desc(),
|
|
in_grad_data,
|
|
rnn.init_h_desc(),
|
|
init_h_grad_data,
|
|
rnn.init_c_desc(),
|
|
init_c_grad_data,
|
|
nullptr,
|
|
nullptr,
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
|
|
PADDLE_ENFORCE_GPU_SUCCESS(
|
|
dynload::cudnnRNNBackwardWeightsEx(handle,
|
|
rnn.rnn_desc(),
|
|
rnn.x_seq_desc(),
|
|
x.data<T>(),
|
|
rnn.init_h_desc(),
|
|
init_h.data<T>(),
|
|
rnn.y_seq_desc(),
|
|
out.data<T>(),
|
|
workspace_data_.data<uint8_t>(),
|
|
workspace_size,
|
|
rnn.weight_desc(),
|
|
weight_grad_data,
|
|
const_cast<uint8_t *>(reserve_data),
|
|
reserve_size));
|
|
#else
|
|
PADDLE_THROW(common::errors::Unavailable(
|
|
"The padded input of rnn is supported by cudnnRNNBackwardDataEx, "
|
|
"cudnnRNNBackwardWeightsEx, but it only works when the version "
|
|
"of cudnn is larger than 7.2.1"));
|
|
#endif
|
|
}
|
|
|
|
#endif // end CUDNN_VERSION >= 90000
|
|
}
|
|
|
|
} // namespace phi
|
|
|
|
#ifdef PADDLE_WITH_HIP
|
|
PD_REGISTER_KERNEL(
|
|
cudnn_lstm_grad, GPU, ALL_LAYOUT, phi::CudnnLSTMGradKernel, float) {}
|
|
#else
|
|
PD_REGISTER_KERNEL(
|
|
cudnn_lstm_grad, GPU, ALL_LAYOUT, phi::CudnnLSTMGradKernel, float, double) {
|
|
}
|
|
#endif
|