Skip to content

N
NiuTrans.Tensor
Commit 28fd2d94 by xiaotong
Options

new code

parent 992ee9e9
正在显示 包含 523 行增加 和 64 行删除
source/network/XBackwardLoss.cpp 0 → 100644
/* NiuTrans.Tensor - an open-source tensor library
* Copyright (C) 2018, Natural Language Processing Lab, Northestern University.
* 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.
*/
/*
* $Created by: XIAO Tong (xiaotong@mail.neu.edu.cn) 2018-07-17
*/
#include "XBackwardLoss.h"
#include "../tensor/XName.h"
#include "../tensor/function/HardTanH.h"
#include "../tensor/function/LogSoftmax.h"
namespace nts{
/*
compute dE/dx for a given function y = f(x)
>> gold - gold standard to measure error (or loss)
>> y - output of the function
>> x - input of the function
>> dedy - dE/dy
>> dedx - dE/dx
>> funcID - id of the function f
>> params - parameters of the function
>> lossName - name of the loss, e.g., cross entropy
*/
void XLossGrad::Compute(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx,
int funcID, void * params,
LOSS_FUNCTION_NAME lossName)
{
CheckNTErrors(gold && y && x, "Empty input tensors!");
CheckNTErrors(dedx, "Empty gradient tensors!");
CheckNTErrors((funcID & FUNCTION_BASE) != 0, "Illegal function id");
if(funcID == FUNC_HARDTANH){
_HardTanHBackward(gold, y, x, dedy, dedx, lossName);
}
else if(funcID == FUNC_LOGSOFTMAX){
int leadDim = *(int*)params;
_LogSoftmaxBackward(gold, y, x, dedy, dedx, leadDim, lossName);
}
else{
ShowNTErrors("wrong function found when call the backward process!");
}
}
/*
compute dE/dy for variable y and error(loss) function E
>> gold - gold standard to measure error (or loss)
>> y - output of the function
>> dedy - dE/dy
>> lossName - name of the loss, e.g., cross entropy
*/
void XLossGrad::Compute(XTensor * gold, XTensor * y,
XTensor * dedy,
LOSS_FUNCTION_NAME lossName)
{
LossBackward(dedy, gold, y, lossName);
}
}
\ No newline at end of file
source/network/XBackwardLoss.h 0 → 100644
/* NiuTrans.Tensor - an open-source tensor library
* Copyright (C) 2018, Natural Language Processing Lab, Northestern University.
* 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.
*/
/*
* $Created by: XIAO Tong (xiaotong@mail.neu.edu.cn) 2018-07-17
* My students worked all night to prepare a submission to CWMT. Good luck
* to them!
*/
#include "../tensor/XTensor.h"
#include "../tensor/function/FHeader.h"
#ifndef __XBACKWARDLOSS_H__
#define __XBACKWARDLOSS_H__
namespace nts{
/* this class computes the gradient (of a output node)
with respect to the loss */
class XLossGrad
{
public:
/* compute dE/dx for a given function y = f(x) */
void Compute(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx,
int funcID, void * params,
LOSS_FUNCTION_NAME lossName);
/* compute dE/dy for variable y and error(loss) function E */
void Compute(XTensor * gold, XTensor * y,
XTensor * dedy,
LOSS_FUNCTION_NAME lossName);
};
}
#endif
\ No newline at end of file
source/network/XBackwardNode.cpp 0 → 100644
/* NiuTrans.Tensor - an open-source tensor library
* Copyright (C) 2018, Natural Language Processing Lab, Northestern University.
* 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.
*/
/*
* $Created by: XIAO Tong (xiaotong@mail.neu.edu.cn) 2018-07-17
*/
#include "XBackwardNode.h"
#include "../tensor/core/CHeader.h"
#include "../tensor/XName.h"
namespace nts{
/* make gradient tensor for a node */
void XNoder::MakeGrad(XTensor * node)
{
if(node == NULL)
return;
if(!XTensor::IsIdentical(node, node->grad)){
delete node->grad;
node->grad = NewTensor(node);
}
}
/* the node is a leaf node (intput) or not */
bool XNoder::IsLeaf(XTensor * node)
{
if(node == NULL)
return false;
if(node->income.tailNum == 0)
return true;
else
return false;
}
/* the node is a root node (output) or not */
bool XNoder::IsRoot(XTensor * node)
{
if(node == NULL)
return false;
if(node->outgo.tailNum == 0)
return true;
else
return false;
}
/* the node keeps the gradinent or not */
bool XNoder::IsGrad(XTensor * node)
{
if(node == NULL)
return false;
if(node->isGrad)
return true;
else
return false;
}
/*
compute dE/dx of a node
Note that input of the node could be arbitrary tensors
>> node - node of the network
*/
void XNodeGrad::Compute(XTensor * node)
{
if(node == NULL || node->visitMark == NODE_FINISHED)
return;
if(!XNoder::IsLeaf(node)){
}
node->visitMark = NODE_FINISHED;
}
/* indicates whether the node is for a math operation */
bool XNodeGrad::IsMathOP(XTensor * node)
{
XLink &income = node->income;
return (income.typeID & MATH_BASE) != 0;
}
/* compute dE/dx as a math operation, e.g., sum, multiply ... */
void XNodeGrad::ComputeMath(XTensor * node)
{
CheckNTErrors(node->grad != NULL, "No gradient found!");
XLink &income = node->income;
int operID = income.typeID;
/* c = a + b * \beta
dE/da = dE/dc
dE/db = dE/dc * \beta */
if(operID == MATH_SUM){
CheckNTErrors(income.tailNum == 2, "Wrong input tensor number for SUM!");
XTensor * a = income.tails[0];
XTensor * b = income.tails[1];
DTYPE beta = income.GetParam(0);
XNoder::MakeGrad(a);
XNoder::MakeGrad(b);
_CopyValues(node->grad, a->grad);
if(beta != 1.0F)
_ScaleAndShift(node->grad, a->grad, beta);
else
_CopyValues(node->grad, b->grad);
}
/* c = a * b
dE/da = dE/dc * b
dE/db = dE/dc * a */
else if(operID == MATH_MULTIPLY){
CheckNTErrors(income.tailNum == 2, "Wrong input tensor number for MULTIPLY!");
XTensor * a = income.tails[0];
XTensor * b = income.tails[1];
int leadDim = income.GetParamInt(0);
XNoder::MakeGrad(a);
XNoder::MakeGrad(b);
CheckNTErrors(XTensor::IsIdentical(a, b), "Wrong sized input tensors!");
_Multiply(node->grad, b, a->grad);
_Multiply(node->grad, a, b->grad);
}
else{
ShowNTErrors("TODO!");
}
}
}
\ No newline at end of file
source/network/XBackwardNode.h 0 → 100644
/* NiuTrans.Tensor - an open-source tensor library
* Copyright (C) 2018, Natural Language Processing Lab, Northestern University.
* 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.
*/
/*
* $Created by: XIAO Tong (xiaotong@mail.neu.edu.cn) 2018-07-17
*/
#include "../tensor/XTensor.h"
#include "../tensor/function/FHeader.h"
#ifndef __XBACKWARDNODE_H__
#define __XBACKWARDNODE_H__
namespace nts{
#define NODE_UNFINISHED 0
#define NODE_DOING 1
#define NODE_FINISHED 2
/* node management */
class XNoder
{
public:
/* make gradient tensor for a node */
static
void MakeGrad(XTensor * node);
/* the node is a leaf node (intput) or not */
static
bool IsLeaf(XTensor * node);
/* the node is a root node (output) or not */
static
bool IsRoot(XTensor * node);
/* the node keeps the gradinent or not */
static
bool IsGrad(XTensor * node);
};
/* this class computes the gradient for each node in the network */
class XNodeGrad
{
public:
/* compute dE/dx of a node */
void Compute(XTensor * node);
/* indicates whether the node is for a math operation */
bool IsMathOP(XTensor * node);
/* compute dE/dx as a math operation, e.g., sum, multiply ... */
void ComputeMath(XTensor * node);
};
}
#endif
\ No newline at end of file
source/network/XNet.cpp
...@@ -20,6 +20,9 @@ ...@@ -20,6 +20,9 @@
*/ */
#include "XNet.h" #include "XNet.h"
#include "XBackwardLoss.h"
#include "XBackwardNode.h"
#include "../tensor/XName.h"
namespace nts{ namespace nts{
...@@ -87,6 +90,44 @@ with a number of root nodes ...@@ -87,6 +90,44 @@ with a number of root nodes
void XNet::Backward(XList &roots, XList &golds, LOSS_FUNCTION_NAME loss) void XNet::Backward(XList &roots, XList &golds, LOSS_FUNCTION_NAME loss)
{ {
Traverse(roots); Traverse(roots);
for(int i = 0; i < nodes.count; i++){
XTensor * node = (XTensor*)nodes.Get(i);
node->visitMark = NODE_UNFINISHED;
}
XLossGrad lossGrad;
/* we start with the gradient with respect to the loss for output layers */
for(int i = 0; i < roots.count; i++){
XTensor * root = (XTensor*)roots.Get(i);
XTensor * gold = (XTensor*)golds.Get(i);
XLink &income = root->income;
int funcID = income.typeID;
void * params = income.params;
/* we compute dE/dx if the output is generated by an activation function y = f(x).
Note that we do not need to obtain dE/dy here because it is no use in the
folloing process of back-propagation */
if(income.tailNum == 1 && (funcID & FUNCTION_BASE)){
XTensor * x = income.tails[0];
XNoder::MakeGrad(x);
lossGrad.Compute(gold, root, x, NULL, x->grad, funcID, params, loss);
root->visitMark = NODE_FINISHED;
}
/* we compuate dE/dy (y is the output) if no predefined activation function is used */
else{
XNoder::MakeGrad(root);
lossGrad.Compute(gold, root, root->grad, loss);
}
}
/* back-propagation from output to input */
for(int i = nodes.count - 1; i >= 0; i--){
XTensor * node = (XTensor*)nodes.Get(i);
if(node->visitMark == NODE_FINISHED)
continue;
}
} }
/* /*
...@@ -115,6 +156,15 @@ void XNet::Traverse(XList &roots) ...@@ -115,6 +156,15 @@ void XNet::Traverse(XList &roots)
for (int i = 0; i < roots.count; i++) for (int i = 0; i < roots.count; i++)
TarjanVisit((XTensor*)roots.Get(i), nodes, id); TarjanVisit((XTensor*)roots.Get(i), nodes, id);
for(int i = 0; i < nodes.count; i++){
XTensor * node = (XTensor*)nodes.Get(i);
if(XNoder::IsRoot(node))
outputs.Add(node);
if(XNoder::IsLeaf(node))
inputs.Add(node);
if(XNoder::IsGrad(node))
gradNodes.Add(node);
}
} }
/* /*
......
source/tensor/Main.cpp
...@@ -45,11 +45,10 @@ int main( int argc, const char ** argv ) ...@@ -45,11 +45,10 @@ int main( int argc, const char ** argv )
//_CrtSetBreakAlloc(123); //_CrtSetBreakAlloc(123);
/* a tiny test */ /* a tiny test */
if(true)
SmallTest(); SmallTest();
//_CrtDumpMemoryLeaks(); //_CrtDumpMemoryLeaks();
return 0; //return 0;
if(argc > 1 && !strcmp(argv[1], "-test")) if(argc > 1 && !strcmp(argv[1], "-test"))
Test(); Test();
......
source/tensor/XLink.cpp
...@@ -25,7 +25,7 @@ ...@@ -25,7 +25,7 @@
namespace nts{ // namespace nts(NiuTrans.Tensor) namespace nts{ // namespace nts(NiuTrans.Tensor)
int XLink::paramSize = 64; int XLink::paramSize = PARAM_UNTI_SIZE;
/* constuctor */ /* constuctor */
XLink::XLink() XLink::XLink()
...@@ -235,6 +235,26 @@ void XLink::AddParam(void * param, int size) ...@@ -235,6 +235,26 @@ void XLink::AddParam(void * param, int size)
} }
/* /*
get a paramter in default type
>> i - id the of the parameter
*/
DTYPE XLink::GetParam(int i)
{
char * p = (char*)params + i * paramSize;
return *(DTYPE*)p;
}
/*
get a paramter in integer
>> i - id the of the parameter
*/
int XLink::GetParamInt(int i)
{
char * p = (char*)params + i * paramSize;
return *(int*)p;
}
/*
create a hyperedge with two input tensors and a output tensor create a hyperedge with two input tensors and a output tensor
>> t1 - a tail tensor >> t1 - a tail tensor
>> t2 - another tail tensor >> t2 - another tail tensor
......
source/tensor/XLink.h
...@@ -34,6 +34,7 @@ namespace nts{ // namespace nts(NiuTrans.Tensor) ...@@ -34,6 +34,7 @@ namespace nts{ // namespace nts(NiuTrans.Tensor)
struct XTensor; struct XTensor;
#define MAX_OP_NAME_LENGTH 16 #define MAX_OP_NAME_LENGTH 16
#define PARAM_UNTI_SIZE 64
/* /*
This defines the link among tensors in networks. XLink can be This defines the link among tensors in networks. XLink can be
...@@ -115,12 +116,18 @@ struct XLink ...@@ -115,12 +116,18 @@ struct XLink
/* add two tails in one time */ /* add two tails in one time */
void AddTwoTails(XTensor * t1, XTensor * t2); void AddTwoTails(XTensor * t1, XTensor * t2);
/* add a integer parameter */ /* add a parameter in default type */
void AddParam(DTYPE param); void AddParam(DTYPE param);
/* add a integer parameter */ /* add a parameter */
void AddParam(void * param, int size); void AddParam(void * param, int size);
/* get a paramter in default type */
DTYPE GetParam(int i);
/* get a paramter in integer */
int GetParamInt(int i);
/* create a hyper edge with two input tensors and a output tensor */ /* create a hyper edge with two input tensors and a output tensor */
static static
void MakeLink(const XTensor * t1, const XTensor * t2, XTensor * h, int id); void MakeLink(const XTensor * t1, const XTensor * t2, XTensor * h, int id);
......
source/tensor/XName.cpp
...@@ -26,7 +26,7 @@ namespace nts { // namespace nts(NiuTrans.Tensor) ...@@ -26,7 +26,7 @@ namespace nts { // namespace nts(NiuTrans.Tensor)
/* get operator name */ /* get operator name */
const char * GetOPName(int type) const char * GetOPName(int type)
{ {
if((type & MATH_ARITHMETIC) != 0){ if((type & MATH_BASE) != 0){
if(type == MATH_ABSOLUTE) if(type == MATH_ABSOLUTE)
return "M_ABSOLUTE"; return "M_ABSOLUTE";
else if(type == MATH_MATRIXMUL) else if(type == MATH_MATRIXMUL)
...@@ -49,12 +49,6 @@ const char * GetOPName(int type) ...@@ -49,12 +49,6 @@ const char * GetOPName(int type)
return "M_POWER"; return "M_POWER";
else if(type == MATH_SCALEANDSHIFT) else if(type == MATH_SCALEANDSHIFT)
return "M_SCALEANDSHIFT"; return "M_SCALEANDSHIFT";
else if(type == GETANDSET_SELECT)
return "G_SELECT";
else if(type == MOVEMENT_COPYINDEXED)
return "M_COPYINDEXED";
else if(type == MOVEMENT_COPYVALUES)
return "M_COPYVALUES";
else if(type == REDUCE_REDUCEMAX) else if(type == REDUCE_REDUCEMAX)
return "R_REDUCEMAX"; return "R_REDUCEMAX";
else if(type == REDUCE_REDUCEMEAN) else if(type == REDUCE_REDUCEMEAN)
...@@ -65,6 +59,14 @@ const char * GetOPName(int type) ...@@ -65,6 +59,14 @@ const char * GetOPName(int type)
return "R_REDUCESUMSQUARED"; return "R_REDUCESUMSQUARED";
else if(type == REDUCE_REDUCEVARIANCE) else if(type == REDUCE_REDUCEVARIANCE)
return "R_REDUCEVARIANCE"; return "R_REDUCEVARIANCE";
}
else if((type & DATA_BASE) != 0){
if(type == GETANDSET_SELECT)
return "G_SELECT";
else if(type == MOVEMENT_COPYINDEXED)
return "M_COPYINDEXED";
else if(type == MOVEMENT_COPYVALUES)
return "M_COPYVALUES";
else if(type == SHAPE_CONCATENATE) else if(type == SHAPE_CONCATENATE)
return "S_CONCATENATE"; return "S_CONCATENATE";
else if(type == SHAPE_MERGE) else if(type == SHAPE_MERGE)
......
source/tensor/XName.h
...@@ -28,8 +28,9 @@ ...@@ -28,8 +28,9 @@
namespace nts { // namespace nts(NiuTrans.Tensor) namespace nts { // namespace nts(NiuTrans.Tensor)
#define MATH_ARITHMETIC 0x00001000 /* math operations */
#define MATH_ABSOLUTE MATH_ARITHMETIC + 1 #define MATH_BASE 0x00001000
#define MATH_ABSOLUTE MATH_BASE + 1
#define MATH_MATRIXMUL MATH_ABSOLUTE + 1 #define MATH_MATRIXMUL MATH_ABSOLUTE + 1
#define MATH_MATRIXMULBATCHED MATH_MATRIXMUL + 1 #define MATH_MATRIXMULBATCHED MATH_MATRIXMUL + 1
#define MATH_MULTIPLY MATH_MATRIXMULBATCHED + 1 #define MATH_MULTIPLY MATH_MATRIXMULBATCHED + 1
...@@ -42,20 +43,22 @@ namespace nts { // namespace nts(NiuTrans.Tensor) ...@@ -42,20 +43,22 @@ namespace nts { // namespace nts(NiuTrans.Tensor)
#define MATH_POWER MATH_NORMALIZE + 1 #define MATH_POWER MATH_NORMALIZE + 1
#define MATH_SCALEANDSHIFT MATH_POWER + 1 #define MATH_SCALEANDSHIFT MATH_POWER + 1
#define GETANDSET MATH_SCALEANDSHIFT + 1 #define REDUCE MATH_SCALEANDSHIFT + 1
#define GETANDSET_SELECT GETANDSET + 1
#define MOVEMENT GETANDSET_SELECT + 1
#define MOVEMENT_COPYINDEXED MOVEMENT + 1
#define MOVEMENT_COPYVALUES MOVEMENT_COPYINDEXED + 1
#define REDUCE MOVEMENT_COPYVALUES + 1
#define REDUCE_REDUCEMAX REDUCE + 1 #define REDUCE_REDUCEMAX REDUCE + 1
#define REDUCE_REDUCEMEAN REDUCE_REDUCEMAX + 1 #define REDUCE_REDUCEMEAN REDUCE_REDUCEMAX + 1
#define REDUCE_REDUCESUM REDUCE_REDUCEMEAN + 1 #define REDUCE_REDUCESUM REDUCE_REDUCEMEAN + 1
#define REDUCE_REDUCESUMSQUARED REDUCE_REDUCESUM + 1 #define REDUCE_REDUCESUMSQUARED REDUCE_REDUCESUM + 1
#define REDUCE_REDUCEVARIANCE REDUCE_REDUCESUMSQUARED + 1 #define REDUCE_REDUCEVARIANCE REDUCE_REDUCESUMSQUARED + 1
/* data and shape related operations */
#define DATA_BASE MATH_BASE * 2
#define GETANDSET DATA_BASE + 1
#define GETANDSET_SELECT GETANDSET + 1
#define MOVEMENT GETANDSET_SELECT + 1
#define MOVEMENT_COPYINDEXED MOVEMENT + 1
#define MOVEMENT_COPYVALUES MOVEMENT_COPYINDEXED + 1
#define SHAPE REDUCE_REDUCEVARIANCE + 1 #define SHAPE REDUCE_REDUCEVARIANCE + 1
#define SHAPE_CONCATENATE SHAPE + 1 #define SHAPE_CONCATENATE SHAPE + 1
#define SHAPE_MERGE SHAPE_CONCATENATE + 1 #define SHAPE_MERGE SHAPE_CONCATENATE + 1
...@@ -64,6 +67,15 @@ namespace nts { // namespace nts(NiuTrans.Tensor) ...@@ -64,6 +67,15 @@ namespace nts { // namespace nts(NiuTrans.Tensor)
#define SHAPE_TRANSPOSE SHAPE_SPLIT + 1 #define SHAPE_TRANSPOSE SHAPE_SPLIT + 1
#define SHAPE_UNSQUEEZE SHAPE_TRANSPOSE + 1 #define SHAPE_UNSQUEEZE SHAPE_TRANSPOSE + 1
/* activation functions */
#define FUNCTION_BASE DATA_BASE * 2
#define FUNC_HARDTANH FUNCTION_BASE + 1
#define FUNC_IDENTITY FUNC_HARDTANH + 1
#define FUNC_LOGSOFTMAX FUNC_IDENTITY + 1
#define FUNC_RECTIFY FUNC_LOGSOFTMAX + 1
#define FUNC_SIGMOID FUNC_RECTIFY + 1
#define FUNC_SOFTMAX FUNC_SIGMOID + 1
/* get operator name */ /* get operator name */
const char * GetOPName(int type); const char * GetOPName(int type);
......
source/tensor/XTensor.cpp
...@@ -352,6 +352,9 @@ judge whether the two matrices are in the same type and size ...@@ -352,6 +352,9 @@ judge whether the two matrices are in the same type and size
*/ */
bool XTensor::IsIdentical(const XTensor * a, const XTensor * b) bool XTensor::IsIdentical(const XTensor * a, const XTensor * b)
{ {
if(a == NULL || b == NULL)
return false;
if(a->order != b->order) if(a->order != b->order)
return false; return false;
......
source/tensor/core/arithmetic/Multiply.cpp
...@@ -142,16 +142,15 @@ void _MultiplyMe(XTensor * a, const XTensor * b, DTYPE alpha, int leadingDim) ...@@ -142,16 +142,15 @@ void _MultiplyMe(XTensor * a, const XTensor * b, DTYPE alpha, int leadingDim)
element-wise product of two tensors (return a XTensor structure) element-wise product of two tensors (return a XTensor structure)
make a new tensor c to keep the result and return it make a new tensor c to keep the result and return it
c(i) = a(i)*b(i) + \alpha * c(i) c(i) = a(i)*b(i)
where i is the index of the item where i is the index of the item
>> a - tensor a >> a - tensor a
>> b - tensor b >> b - tensor b
>> alpha - the coefficient
>> leadingDim - the dimension along which we perform broadcasting >> leadingDim - the dimension along which we perform broadcasting
<< return - the product of the tensors << return - the product of the tensors
*/ */
XTensor Multiply(const XTensor &a, const XTensor &b, DTYPE alpha, int leadingDim) XTensor Multiply(const XTensor &a, const XTensor &b, int leadingDim)
{ {
CheckNTErrors(a.dimSize[leadingDim] == b.dimSize[leadingDim], "TODO!"); CheckNTErrors(a.dimSize[leadingDim] == b.dimSize[leadingDim], "TODO!");
...@@ -159,11 +158,10 @@ XTensor Multiply(const XTensor &a, const XTensor &b, DTYPE alpha, int leadingDim ...@@ -159,11 +158,10 @@ XTensor Multiply(const XTensor &a, const XTensor &b, DTYPE alpha, int leadingDim
c.SetTMP(); c.SetTMP();
/* call _Multiply function */ /* call _Multiply function */
_Multiply(&a, &b, &c, alpha, leadingDim); _Multiply(&a, &b, &c, 0, leadingDim);
/* tensor connections */ /* tensor connections */
XLink::MakeLink(&a, &b, &c, MATH_MULTIPLY); XLink::MakeLink(&a, &b, &c, MATH_MULTIPLY);
XLink::AddParamToHead(&c, alpha);
XLink::AddParamToHeadInt(&c, leadingDim); XLink::AddParamToHeadInt(&c, leadingDim);
return c; return c;
......
source/tensor/core/arithmetic/Multiply.h
...@@ -44,10 +44,10 @@ void _MultiplyMe(XTensor * a, const XTensor * b, DTYPE alpha = 0, int leadingDim ...@@ -44,10 +44,10 @@ void _MultiplyMe(XTensor * a, const XTensor * b, DTYPE alpha = 0, int leadingDim
/* /*
element-wise product of two tensors (return a XTensor structure) element-wise product of two tensors (return a XTensor structure)
make a new tensor to keep the result and return it make a new tensor to keep the result and return it
c(i) = a(i)*b(i) + \alpha * c(i) c(i) = a(i)*b(i)
where i is the index of the element where i is the index of the element
*/ */
XTensor Multiply(const XTensor &a, const XTensor &b, DTYPE alpha = 0, int leadingDim = 0); XTensor Multiply(const XTensor &a, const XTensor &b, int leadingDim = 0);
} // namespace nts(NiuTrans.Tensor) } // namespace nts(NiuTrans.Tensor)
......
source/tensor/function/HardTanH.cpp
...@@ -20,6 +20,7 @@ ...@@ -20,6 +20,7 @@
*/ */
#include <stdlib.h> #include <stdlib.h>
#include "../XName.h"
#include "HardTanH.h" #include "HardTanH.h"
#include "HardTanH.cuh" #include "HardTanH.cuh"
...@@ -59,6 +60,26 @@ void _HardTanH(const XTensor * x, XTensor * y) ...@@ -59,6 +60,26 @@ void _HardTanH(const XTensor * x, XTensor * y)
} }
/* /*
hard tanh function (return a structure)
y = 1 if x > 1
x if -1 <= x <= 1
-1 if x < -1
>> x - input tensor
<< return - y
*/
XTensor HardTanH(const XTensor &x)
{
XTensor y(&x);
y.SetTMP();
_HardTanH(&x, &y);
XLink::MakeLink(&x, NULL, &y, FUNC_HARDTANH);
return y;
}
/*
backward computation backward computation
dE/dx = dE/dy * dy/dx dE/dx = dE/dy * dy/dx
...@@ -77,7 +98,7 @@ hard tanh: y = 1 if x > 1 ...@@ -77,7 +98,7 @@ hard tanh: y = 1 if x > 1
>> dedx - dE/dx >> dedx - dE/dx
>> lossName - type of loss function, e.g., cross entropy >> lossName - type of loss function, e.g., cross entropy
*/ */
void HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x, void _HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
LOSS_FUNCTION_NAME lossName) LOSS_FUNCTION_NAME lossName)
{ {
...@@ -86,7 +107,7 @@ void HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x, ...@@ -86,7 +107,7 @@ void HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
#ifdef USE_CUDA #ifdef USE_CUDA
if(x->devID >= 0 || y->devID >= 0){ if(x->devID >= 0 || y->devID >= 0){
CudaHardTanHBackward(gold, y, x, dedy, dedx, lossName); _CudaHardTanHBackward(gold, y, x, dedy, dedx, lossName);
return; return;
} }
#endif #endif
......
source/tensor/function/HardTanH.cu
...@@ -129,7 +129,7 @@ hard tanh: y = 1 if x > 1 ...@@ -129,7 +129,7 @@ hard tanh: y = 1 if x > 1
>> dedx - dE/dx >> dedx - dE/dx
>> lossName - type of loss function, e.g., cross entropy >> lossName - type of loss function, e.g., cross entropy
*/ */
void CudaHardTanHBackward(XTensor * gold, XTensor * y, XTensor * x, void _CudaHardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
LOSS_FUNCTION_NAME lossName) LOSS_FUNCTION_NAME lossName)
{ {
......
source/tensor/function/HardTanH.cuh
...@@ -40,7 +40,7 @@ void _CudaHardTanH(const XTensor * input, XTensor * output); ...@@ -40,7 +40,7 @@ void _CudaHardTanH(const XTensor * input, XTensor * output);
/* de/dx (Cuda version) */ /* de/dx (Cuda version) */
extern "C" extern "C"
void CudaHardTanHBackward(XTensor * gold, XTensor * y, XTensor * x, void _CudaHardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
LOSS_FUNCTION_NAME lossName); LOSS_FUNCTION_NAME lossName);
......
source/tensor/function/HardTanH.h
...@@ -33,12 +33,13 @@ y = 1 if x > 1 ...@@ -33,12 +33,13 @@ y = 1 if x > 1
x if -1 <= x <= 1 x if -1 <= x <= 1
-1 if x < -1 -1 if x < -1
*/ */
extern "C"
void _HardTanH(const XTensor * x, XTensor * y); void _HardTanH(const XTensor * x, XTensor * y);
/* hard tanh function (return a structure) */
XTensor HardTanH(const XTensor &x);
/* de/dx */ /* de/dx */
extern "C" void _HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
void HardTanHBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
LOSS_FUNCTION_NAME lossName); LOSS_FUNCTION_NAME lossName);
......
source/tensor/function/LogSoftmax.cpp
...@@ -122,7 +122,7 @@ void _LogSoftmax(const XTensor * x, XTensor * y, int leadDim) ...@@ -122,7 +122,7 @@ void _LogSoftmax(const XTensor * x, XTensor * y, int leadDim)
blockMax->data = mp; blockMax->data = mp;
blockSum->data = sp; blockSum->data = sp;
#ifdef USE_CUDA #ifdef USE_CUDA
CudaLogSoftmaxSumMax(blockx, blocky, leadDim, blockSum, blockMax); _CudaLogSoftmaxSumMax(blockx, blocky, leadDim, blockSum, blockMax);
#else #else
ShowNTErrors("Please specify USE_CUDA and recompile the code!"); ShowNTErrors("Please specify USE_CUDA and recompile the code!");
#endif #endif
...@@ -223,7 +223,7 @@ better numerical stability. ...@@ -223,7 +223,7 @@ better numerical stability.
>> lossName - type of loss function, e.g., cross entropy >> lossName - type of loss function, e.g., cross entropy
>> leadDim - leading dimension (along which we perform reduction) >> leadDim - leading dimension (along which we perform reduction)
*/ */
void LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x, void _LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
int leadDim, int leadDim,
LOSS_FUNCTION_NAME lossName) LOSS_FUNCTION_NAME lossName)
...@@ -234,7 +234,7 @@ void LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x, ...@@ -234,7 +234,7 @@ void LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
int leadDimRDI = y->order - leadDim - 1; int leadDimRDI = y->order - leadDim - 1;
#ifdef USE_CUDA #ifdef USE_CUDA
if (gold->devID >= 0) { if (gold->devID >= 0) {
CudaLogSoftmaxBackward(gold, y, x, dedy, dedx, leadDim, lossName); _CudaLogSoftmaxBackward(gold, y, x, dedy, dedx, leadDim, lossName);
return; return;
} }
#endif #endif
......
source/tensor/function/LogSoftmax.cu
...@@ -139,7 +139,7 @@ log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version) ...@@ -139,7 +139,7 @@ log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version)
>> sum - \sum_{i} e^{x_i} >> sum - \sum_{i} e^{x_i}
>> max - \max_{i} e^{x_i} >> max - \max_{i} e^{x_i}
*/ */
void CudaLogSoftmaxSumMax(XTensor * x, XTensor * y, int leadDim, XTensor * sum, XTensor * max) void _CudaLogSoftmaxSumMax(XTensor * x, XTensor * y, int leadDim, XTensor * sum, XTensor * max)
{ {
CheckNTErrors((x->devID >= 0), "Forward computation of log softmax must be run on GPUs."); CheckNTErrors((x->devID >= 0), "Forward computation of log softmax must be run on GPUs.");
CheckNTErrors((x->devID == y->devID), "Input tensors must be on the same GPU."); CheckNTErrors((x->devID == y->devID), "Input tensors must be on the same GPU.");
...@@ -353,7 +353,7 @@ better numerical stability. ...@@ -353,7 +353,7 @@ better numerical stability.
>> lossName - type of loss function, e.g., cross entropy >> lossName - type of loss function, e.g., cross entropy
>> leadDim - leading dimension (along which we perform reduction) >> leadDim - leading dimension (along which we perform reduction)
*/ */
void CudaLogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x, void _CudaLogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
int leadDim, int leadDim,
LOSS_FUNCTION_NAME lossName) LOSS_FUNCTION_NAME lossName)
......
source/tensor/function/LogSoftmax.cuh
...@@ -30,16 +30,13 @@ namespace nts { // namespace nts(NiuTrans.Tensor) ...@@ -30,16 +30,13 @@ namespace nts { // namespace nts(NiuTrans.Tensor)
#ifdef USE_CUDA #ifdef USE_CUDA
/* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version) */ /* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version) */
extern "C"
void _CudaLogSoftmax(const XTensor * input, XTensor * output, int leadDim); void _CudaLogSoftmax(const XTensor * input, XTensor * output, int leadDim);
/* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version) */ /* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) (Cuda version) */
extern "C" void _CudaLogSoftmaxSumMax(XTensor * x, XTensor * y, int leadDim, XTensor * sum, XTensor * max);
void CudaLogSoftmaxSumMax(XTensor * x, XTensor * y, int leadDim, XTensor * sum, XTensor * max);
/* de/dx (Cuda version) */ /* de/dx (Cuda version) */
extern "C" void _CudaLogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
void CudaLogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
int leadDim, int leadDim,
LOSS_FUNCTION_NAME lossName); LOSS_FUNCTION_NAME lossName);
......
source/tensor/function/LogSoftmax.h
...@@ -28,12 +28,10 @@ ...@@ -28,12 +28,10 @@
namespace nts{ // namespace nts(NiuTrans.Tensor) namespace nts{ // namespace nts(NiuTrans.Tensor)
/* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) */ /* log scale softmax y = log(e^x / \sum_{i} e^{x_i}) */
extern "C"
void _LogSoftmax(const XTensor * x, XTensor * y, int leadDim); void _LogSoftmax(const XTensor * x, XTensor * y, int leadDim);
/* de/dx */ /* de/dx */
extern "C" void _LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
void LogSoftmaxBackward(XTensor * gold, XTensor * y, XTensor * x,
XTensor * dedy, XTensor * dedx, XTensor * dedy, XTensor * dedx,
int leadDim, int leadDim,
LOSS_FUNCTION_NAME lossName); LOSS_FUNCTION_NAME lossName);
......
source/tensor/function/Loss.cu
...@@ -405,7 +405,7 @@ void CudaLossBackward(XTensor * dedy, XTensor * t, XTensor * y, ...@@ -405,7 +405,7 @@ void CudaLossBackward(XTensor * dedy, XTensor * t, XTensor * y,
loss = sum_{i} (-t_i * log(y_i)), where t and y are distributions loss = sum_{i} (-t_i * log(y_i)), where t and y are distributions
dloss/dy_i = -t_i / y_i dloss/dy_i = -t_i / y_i
*/ */
if(LFName == CROSSENTROPY){ else if(LFName == CROSSENTROPY){
if(t->isSparse){ if(t->isSparse){
ShowNTErrors("TODO!"); ShowNTErrors("TODO!");
} }
...@@ -416,6 +416,9 @@ void CudaLossBackward(XTensor * dedy, XTensor * t, XTensor * y, ...@@ -416,6 +416,9 @@ void CudaLossBackward(XTensor * dedy, XTensor * t, XTensor * y,
KernelLossBackwardCrossEntropyBlock<<<blocks, threads>>>(dedyp, tp, yp, blockSize, tBeg * stride, tLen * stride, y->unitNum); KernelLossBackwardCrossEntropyBlock<<<blocks, threads>>>(dedyp, tp, yp, blockSize, tBeg * stride, tLen * stride, y->unitNum);
} }
} }
else{
ShowNTErrors("TODO");
}
BacktoCudaDev(y->devID, devIDBackup); BacktoCudaDev(y->devID, devIDBackup);
} }
......
source/tensor/sample/fnnlm/FNNLM.cpp
...@@ -773,7 +773,7 @@ void Backward(XTensor inputs[], XTensor &output, XTensor &gold, LOSS_FUNCTION_NA ...@@ -773,7 +773,7 @@ void Backward(XTensor inputs[], XTensor &output, XTensor &gold, LOSS_FUNCTION_NA
/* for y = softmax(s), we get dE/ds /* for y = softmax(s), we get dE/ds
where E is the error function (define by loss) */ where E is the error function (define by loss) */
LogSoftmaxBackward(&gold, &y, &s, NULL, &deds, 1, loss); _LogSoftmaxBackward(&gold, &y, &s, NULL, &deds, 1, loss);
/* for s = x * w, we get /* for s = x * w, we get
dE/w_{i,j} = dE/ds_j * ds/dw_{i,j} dE/w_{i,j} = dE/ds_j * ds/dw_{i,j}
...@@ -818,7 +818,7 @@ void Backward(XTensor inputs[], XTensor &output, XTensor &gold, LOSS_FUNCTION_NA ...@@ -818,7 +818,7 @@ void Backward(XTensor inputs[], XTensor &output, XTensor &gold, LOSS_FUNCTION_NA
/* backpropagation through the activation fucntion: /* backpropagation through the activation fucntion:
dE/ds = dE/dh * dh/ds */ dE/ds = dE/dh * dh/ds */
HardTanHBackward(NULL, &h, &s, &dedh, &deds, NOLOSS); _HardTanHBackward(NULL, &h, &s, &dedh, &deds, NOLOSS);
/* gradient of the weight: dE/dw = x^T * dE/ds */ /* gradient of the weight: dE/dw = x^T * dE/ds */
_MatrixMul(&x, X_TRANS, &deds, X_NOTRANS, &dedw); _MatrixMul(&x, X_TRANS, &deds, X_NOTRANS, &dedw);
......
source/tensor/test/THardTanH.cpp
...@@ -155,7 +155,7 @@ bool TestHardTanH2() ...@@ -155,7 +155,7 @@ bool TestHardTanH2()
_HardTanH(x, y); _HardTanH(x, y);
/* call HardTanHBackward function */ /* call HardTanHBackward function */
HardTanHBackward(gold, y, x, dedy, dedx, SQUAREDERROR); _HardTanHBackward(gold, y, x, dedy, dedx, SQUAREDERROR);
/* check results */ /* check results */
cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F) cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F)
...@@ -184,7 +184,7 @@ bool TestHardTanH2() ...@@ -184,7 +184,7 @@ bool TestHardTanH2()
_HardTanH(xGPU, yGPU); _HardTanH(xGPU, yGPU);
/* call hardtanhbackward function */ /* call hardtanhbackward function */
HardTanHBackward(goldGPU, yGPU, xGPU, dedyGPU, dedxGPU, SQUAREDERROR); _HardTanHBackward(goldGPU, yGPU, xGPU, dedyGPU, dedxGPU, SQUAREDERROR);
/* check results */ /* check results */
gpuTest = y->CheckData(yAnswer, unitNum, 1e-4F) gpuTest = y->CheckData(yAnswer, unitNum, 1e-4F)
......
source/tensor/test/TLogSoftmax.cpp
...@@ -142,7 +142,7 @@ bool TestLogSoftmax2() ...@@ -142,7 +142,7 @@ bool TestLogSoftmax2()
_LogSoftmax(x, y, 1); _LogSoftmax(x, y, 1);
/* call LogSoftmaxBackward function */ /* call LogSoftmaxBackward function */
LogSoftmaxBackward(g, y, x, dedy, dedx, 1, CROSSENTROPY); _LogSoftmaxBackward(g, y, x, dedy, dedx, 1, CROSSENTROPY);
/* check result */ /* check result */
cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F) cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F)
...@@ -170,7 +170,7 @@ bool TestLogSoftmax2() ...@@ -170,7 +170,7 @@ bool TestLogSoftmax2()
_LogSoftmax(xGPU, yGPU, 1); _LogSoftmax(xGPU, yGPU, 1);
/* call LogSoftmaxBackward function */ /* call LogSoftmaxBackward function */
LogSoftmaxBackward(gGPU, yGPU, xGPU, dedyGPU, dedxGPU, 1, CROSSENTROPY); _LogSoftmaxBackward(gGPU, yGPU, xGPU, dedyGPU, dedxGPU, 1, CROSSENTROPY);
/* check result */ /* check result */
gpuTest = yGPU->CheckData(yAnswer, unitNum, 1e-4F) && dedxGPU->CheckData(dedxAnswer, unitNum, 1e-4F); gpuTest = yGPU->CheckData(yAnswer, unitNum, 1e-4F) && dedxGPU->CheckData(dedxAnswer, unitNum, 1e-4F);
...@@ -246,7 +246,7 @@ bool TestLogSoftmax3() ...@@ -246,7 +246,7 @@ bool TestLogSoftmax3()
_LogSoftmax(x, y, 1); _LogSoftmax(x, y, 1);
/* call LogSoftmaxBackward function */ /* call LogSoftmaxBackward function */
LogSoftmaxBackward(g, y, x, dedy, dedx, 1, SQUAREDERROR); _LogSoftmaxBackward(g, y, x, dedy, dedx, 1, SQUAREDERROR);
/* check result */ /* check result */
cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F) cpuTest = y->CheckData(yAnswer, unitNum, 1e-4F)
...@@ -274,7 +274,7 @@ bool TestLogSoftmax3() ...@@ -274,7 +274,7 @@ bool TestLogSoftmax3()
_LogSoftmax(xGPU, yGPU, 1); _LogSoftmax(xGPU, yGPU, 1);
/* call LogSoftmaxBackward function */ /* call LogSoftmaxBackward function */
LogSoftmaxBackward(gGPU, yGPU, xGPU, dedyGPU, dedxGPU, 1, SQUAREDERROR); _LogSoftmaxBackward(gGPU, yGPU, xGPU, dedyGPU, dedxGPU, 1, SQUAREDERROR);
/* check result */ /* check result */
gpuTest = yGPU->CheckData(yAnswer, unitNum, 1e-4F) gpuTest = yGPU->CheckData(yAnswer, unitNum, 1e-4F)
......
source/tensor/test/TMultiply.cpp
...@@ -86,7 +86,7 @@ bool TestMultiply1() ...@@ -86,7 +86,7 @@ bool TestMultiply1()
/* call Multiply function */ /* call Multiply function */
_Multiply(s1, s2, t, 0, 0); _Multiply(s1, s2, t, 0, 0);
_MultiplyMe(tMe, s2, 0, 0); _MultiplyMe(tMe, s2, 0, 0);
tUser = Multiply(*s1, *s2, 0, 0); tUser = Multiply(*s1, *s2, 0);
/* check results */ /* check results */
cpuTest = t->CheckData(answer, tUnitNum) cpuTest = t->CheckData(answer, tUnitNum)
...@@ -112,7 +112,7 @@ bool TestMultiply1() ...@@ -112,7 +112,7 @@ bool TestMultiply1()
/* call Multiply function */ /* call Multiply function */
_Multiply(sGPU1, sGPU2, tGPU, 0, 0); _Multiply(sGPU1, sGPU2, tGPU, 0, 0);
_MultiplyMe(tMeGPU, sGPU2, 0, 0); _MultiplyMe(tMeGPU, sGPU2, 0, 0);
tUserGPU = Multiply(*sGPU1, *sGPU2, 0, 0); tUserGPU = Multiply(*sGPU1, *sGPU2, 0);
/* check results */ /* check results */
gpuTest = tGPU->CheckData(answer, tUnitNum) gpuTest = tGPU->CheckData(answer, tUnitNum)
...@@ -209,7 +209,7 @@ bool TestMultiply2() ...@@ -209,7 +209,7 @@ bool TestMultiply2()
/* call Multiply function */ /* call Multiply function */
_Multiply(s1, s2, t, 0, 0); _Multiply(s1, s2, t, 0, 0);
_MultiplyMe(tMe, s2, 0, 0); _MultiplyMe(tMe, s2, 0, 0);
tUser = Multiply(*s1, *s2, 0, 0); tUser = Multiply(*s1, *s2, 0);
/* check results */ /* check results */
cpuTest = t->CheckData(answer, tUnitNum) cpuTest = t->CheckData(answer, tUnitNum)
...@@ -235,11 +235,12 @@ bool TestMultiply2() ...@@ -235,11 +235,12 @@ bool TestMultiply2()
/* call Multiply function */ /* call Multiply function */
_Multiply(sGPU1, sGPU2, tGPU, 0, 0); _Multiply(sGPU1, sGPU2, tGPU, 0, 0);
_MultiplyMe(tMeGPU, sGPU2, 0, 0); _MultiplyMe(tMeGPU, sGPU2, 0, 0);
tUserGPU = Multiply(*sGPU1, *sGPU2, 0, 0); tUserGPU = Multiply(*sGPU1, *sGPU2, 0);
/* check results */ /* check results */
gpuTest = tGPU->CheckData(answer, tUnitNum) gpuTest = tGPU->CheckData(answer, tUnitNum) &&
&& tMeGPU->CheckData(answer, tUnitNum) && tUserGPU.CheckData(answer, tUnitNum); tMeGPU->CheckData(answer, tUnitNum) &&
tUserGPU.CheckData(answer, tUnitNum);
/* destroy variables */ /* destroy variables */
delete s1; delete s1;
......
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论