118 lines
5.2 KiB
Plaintext
118 lines
5.2 KiB
Plaintext
/*******************************************************************************
|
|
* Copyright (c) 2019 Konduit K.K.
|
|
*
|
|
* This program and the accompanying materials are made available under the
|
|
* terms of the Apache License, Version 2.0 which is available at
|
|
* https://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.
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
******************************************************************************/
|
|
|
|
//
|
|
// @author Oleh Semeniv (oleg.semeniv@gmail.com)
|
|
//
|
|
|
|
#include <system/op_boilerplate.h>
|
|
#include <ops/declarable/helpers/updatersHelpers.h>
|
|
#include <helpers/PointersManager.h>
|
|
#include <math/platformmath.h>
|
|
#include <math/templatemath.h>
|
|
|
|
namespace sd {
|
|
namespace ops {
|
|
namespace helpers {
|
|
|
|
///////////////////////////////////////////////////////////////////
|
|
template<typename T>
|
|
__global__ void adaGradUpdaterCuda(const void* vx, const Nd4jLong* xShapeInfo, const void* vin, const Nd4jLong* inShapeInfo,
|
|
void* vz, const Nd4jLong* zShapeInfo, void* vst, const Nd4jLong* stShapeInfo,
|
|
const T lr, const T epsilon) {
|
|
|
|
const auto x = reinterpret_cast<const T*>(vx);
|
|
const auto init = reinterpret_cast<const T*>(vin);
|
|
|
|
auto up = reinterpret_cast<T*>(vz);
|
|
auto st = reinterpret_cast<T*>(vst);
|
|
|
|
__shared__ bool bEWS, bOrdering, bXZsame, bXInSame, bXStSame;
|
|
__shared__ Nd4jLong xLen;
|
|
|
|
if (threadIdx.x == 0) {
|
|
xLen = shape::length(xShapeInfo);
|
|
|
|
bEWS = 1 == shape::elementWiseStride(xShapeInfo) && 1 == shape::elementWiseStride(zShapeInfo) &&
|
|
1 == shape::elementWiseStride(stShapeInfo) && 1 == shape::elementWiseStride(inShapeInfo);
|
|
bOrdering = shape::order(xShapeInfo) == shape::order(zShapeInfo) && shape::order(xShapeInfo) == shape::order(stShapeInfo) &&
|
|
shape::order(xShapeInfo) == shape::order(inShapeInfo);
|
|
|
|
bXZsame = shape::haveSameShapeAndStrides(xShapeInfo, zShapeInfo);
|
|
bXInSame = shape::haveSameShapeAndStrides(xShapeInfo, inShapeInfo);
|
|
bXStSame = shape::haveSameShapeAndStrides(xShapeInfo, stShapeInfo);
|
|
}
|
|
__syncthreads();
|
|
|
|
int coords[MAX_RANK];
|
|
|
|
for (Nd4jLong i = blockIdx.x * blockDim.x + threadIdx.x; i < xLen; i += gridDim.x * blockDim.x) {
|
|
|
|
auto xOffset = i, zOffset = i, initOffset = i, stOffset = i;
|
|
|
|
if (!bEWS || !bOrdering) {
|
|
|
|
shape::index2coords(i, xShapeInfo, coords);
|
|
xOffset = shape::getOffset(xShapeInfo, coords);
|
|
zOffset = bXZsame ? xOffset : shape::getOffset(zShapeInfo, coords);
|
|
initOffset = bXInSame ? xOffset : shape::getOffset(inShapeInfo, coords);
|
|
stOffset = bXStSame ? xOffset : shape::getOffset(stShapeInfo, coords);
|
|
}
|
|
|
|
st[stOffset] = init[initOffset] + x[xOffset] * x[xOffset];
|
|
up[zOffset] = (lr * x[xOffset]) / (math::nd4j_sqrt<T, T>(st[stOffset]) + epsilon);
|
|
|
|
}
|
|
}
|
|
|
|
///////////////////////////////////////////////////////////////////
|
|
template<typename T>
|
|
linkage void adaGradUpdaterCudaLauncher(const int blocksPerGrid, const int threadsPerBlock, const cudaStream_t* stream,
|
|
const void* vx, const Nd4jLong* xShapeInfo, const void* vin, const Nd4jLong* inShapeInfo,
|
|
void* vz, const Nd4jLong* zShapeInfo, void* vst, const Nd4jLong* stShapeInfo,
|
|
const double dLr, const double dEpsilon) {
|
|
|
|
const T lr = static_cast<T>(dLr);
|
|
const T epsilon = static_cast<T>(dEpsilon);
|
|
|
|
adaGradUpdaterCuda<T> << <blocksPerGrid, threadsPerBlock, 256, * stream >> > (vx, xShapeInfo, vin, inShapeInfo,
|
|
vz, zShapeInfo, vst, stShapeInfo, lr, epsilon);
|
|
}
|
|
|
|
///////////////////////////////////////////////////////////////////
|
|
void updaterAdaGrad(sd::LaunchContext* context, const NDArray& gradient, const NDArray& initState,
|
|
NDArray& update, NDArray& stateH, const double dLr, const double dEpsilon) {
|
|
|
|
PointersManager manager(context, "adaGradUpdater");
|
|
|
|
const int threadsPerBlock = MAX_NUM_THREADS / 4;
|
|
const int blocksPerGrid = (gradient.lengthOf() + threadsPerBlock - 1) / threadsPerBlock;
|
|
|
|
NDArray::prepareSpecialUse({ &update, &stateH }, { &gradient, &initState });
|
|
BUILD_SINGLE_SELECTOR(gradient.dataType(), adaGradUpdaterCudaLauncher, (blocksPerGrid, threadsPerBlock, context->getCudaStream(),
|
|
gradient.getSpecialBuffer(), gradient.getSpecialShapeInfo(),
|
|
initState.getSpecialBuffer(), initState.getSpecialShapeInfo(),
|
|
update.getSpecialBuffer(), update.getSpecialShapeInfo(),
|
|
stateH.getSpecialBuffer(), stateH.getSpecialShapeInfo(), dLr, dEpsilon), FLOAT_TYPES);
|
|
NDArray::registerSpecialUse({ &update, &stateH }, { &gradient, &initState });
|
|
|
|
manager.synchronize();
|
|
}
|
|
|
|
}
|
|
}
|
|
}
|