rename XParamKeeper as XTensorKeeper

source/train/XLeader.cpp

@@ -96,7 +96,7 @@ void XLeader::SetServerModel(XConfig * config, XModel * model, XList * memberMod
 {
     serverModel.Clear();
     for (int i = 0; i < model->paramNum; i++) {
-        XTensor * param = model->params[i].param;
+        XTensor * param = model->params[i].tensor;
         serverModel.AddParam(param);
     }
 
@@ -151,7 +151,7 @@ void XLeader::InitForRun()
 void XLeader::ResetParamGrad()
 {
     for (int i = 0; i < serverModel.paramNum; i++) {
-        XTensor* param = serverModel.params[i].param;
+        XTensor* param = serverModel.params[i].tensor;
         if (param->grad != NULL) {
             param->grad->SetZeroAll();
         }
@@ -161,7 +161,7 @@ void XLeader::ResetParamGrad()
         XWorkerJob * worker = (XWorkerJob*)jworkers[j];
         XModel * model = worker->GetModel();
         for (int i = 0; i < model->paramNum; i++) {
-            XTensor* param = model->params[i].param;
+            XTensor* param = model->params[i].tensor;
             if (param->grad != NULL) {
                 param->grad->SetZeroAll();
             }
@@ -517,25 +517,25 @@ void XLeader::RunUpdate(XConfig * config, XOptimizer * optimizer, const int * ac
     while (1) {
         for (int j = 0; j < serverModel.paramNum; j++) {
 
-            XParamKeeper &paramServer = serverModel.params[j];
+            XTensorKeeper &paramServer = serverModel.params[j];
 
             /* isGradFinished is true only if the model finishes the computation
             (in another process) */
-            if (paramServer.flag != PARAM_STATE_NOT_READY || !paramServer.param->isGradFinished)
+            if (paramServer.flag != PARAM_STATE_NOT_READY || !paramServer.tensor->isGradFinished)
                 continue;
 
             /* check if all the models (or part of them) are ready */
             for (int i = 0; i < members.count; i++) {
                 XModel * source = (XModel*)members.GetItem(i);
-                XParamKeeper &paramSource = source->params[j];
+                XTensorKeeper &paramSource = source->params[j];
 
                 /* isGradFinished is true only if the model finishes the computation
                 (in another process) */
-                if (paramSource.flag == PARAM_STATE_NOT_READY && paramSource.param->isGradFinished) {
+                if (paramSource.flag == PARAM_STATE_NOT_READY && paramSource.tensor->isGradFinished) {
                     XQueue* jobQueue = (XQueue*)jobQueues.GetItem(j);
 
                     /* data transmit */
-                    collecter->AddJobCollectDataP2P(jobQueue, paramSource.param->grad, paramServer.param->grad);
+                    collecter->AddJobCollectDataP2P(jobQueue, paramSource.tensor->grad, paramServer.tensor->grad);
                     collecter->AddJobEnqueueFinished(jobQueue);
 
                     /* reset the flag */

source/train/XModel.cpp

@@ -35,9 +35,9 @@ namespace nts {
 
 
 /* constructor */
-XParamKeeper::XParamKeeper()
+XTensorKeeper::XTensorKeeper()
 {
-    param = NULL;
+    tensor = NULL;
     flag = PARAM_STATE_NOT_READY;
     trainFlag = PARAM_STATE_NOT_READY;
     MUTEX_INIT(accessLock);
@@ -45,7 +45,7 @@ XParamKeeper::XParamKeeper()
 }
 
 /* constructor */
-XParamKeeper::~XParamKeeper()
+XTensorKeeper::~XTensorKeeper()
 {
     MUTEX_DELE(accessLock);
     MUTEX_DELE(trainLock);
@@ -124,14 +124,14 @@ void XModel::AddParam(XTensor* param)
 {
     param->SetVarFlag();
 
-    XParamKeeper * newParams = new XParamKeeper[paramNum + 1];
+    XTensorKeeper * newParams = new XTensorKeeper[paramNum + 1];
 
     for (int i = 0; i < paramNum; i++) {
-        newParams[i].param = params[i].param;
+        newParams[i].tensor = params[i].tensor;
         newParams[i].flag = params[i].flag;
     }
 
-    newParams[paramNum].param = param;
+    newParams[paramNum].tensor = param;
     newParams[paramNum].flag = PARAM_STATE_NOT_READY;
 
     delete[] params;
@@ -143,7 +143,7 @@ void XModel::AddParam(XTensor* param)
 bool XModel::CheckParam()
 {
     for (int i = 0; i < paramNum; i++) {
-        XTensor * param = params[i].param;
+        XTensor * param = params[i].tensor;
         if (!param->isGrad)
             return false;
     }
@@ -191,7 +191,7 @@ void XModel::WaitForUnlockedParams()
 void XModel::RefreshMe()
 {
     for (int i = 0; i < paramNum; i++) {
-        params[i].param->isGradFinished = false;
+        params[i].tensor->isGradFinished = false;
         params[i].flag = PARAM_STATE_NOT_READY;
         params[i].trainFlag = PARAM_STATE_NOT_READY;
     }

source/train/XModel.h

@@ -50,12 +50,12 @@ enum PARAM_STATE { PARAM_STATE_NOT_READY,
                    PARAM_STATE_COLLECTED, 
                    PARAM_STATE_UPDATED };
 
-/* parameter keeper */
-class XParamKeeper
+/* tensor keeper */
+class XTensorKeeper
 {
 public:
     /* the parameter */
-    XTensor * param;
+    XTensor * tensor;
 
     /* the parameter state */
     PARAM_STATE flag;
@@ -73,11 +73,10 @@ public:
 
 public:
     /* constructor */
-    XParamKeeper();
+    XTensorKeeper();
 
     /* constructor */
-    ~XParamKeeper();
-
+    ~XTensorKeeper();
 };
 
 /* a model template for training */
@@ -89,7 +88,7 @@ protected:
 
 public:
     /* the list of model parameters */
-    XParamKeeper * params;
+    XTensorKeeper * params;
 
     /* parameter number */
     int paramNum;

source/train/XWorkerBroadcast.cpp

@@ -64,7 +64,7 @@ void XWorkerBroadcast::BroadcastData(XModel * source, XList * targetList, int pi
         XModel * target = (XModel*)targetList->GetItem(i);
 
         /* data transmit */
-        BroadcastP2P(source->params[pid].param, target->params[pid].param);
+        BroadcastP2P(source->params[pid].tensor, target->params[pid].tensor);
 
         /* update the flag */
         target->params[pid].flag = PARAM_STATE_UPDATED;

source/train/XWorkerUpdate.cpp

@@ -64,7 +64,7 @@ void XWorkerUpdate::UpdateParameter(XModel * server, int pid,
 
     CheckNTErrors(server->params[pid].flag == PARAM_STATE_COLLECTED, "The state of the parameter is wrong!");
 
-    XTensor * param = server->params[pid].param;
+    XTensor * param = server->params[pid].tensor;
     XTensor * grad = param->grad;
 
     CheckNTErrors(grad != NULL, "No gradient!");