[Unify Tensors PR #8] Merged Tensor into DenseTensor, test=allcases (PaddlePaddle#38914)

* Merged LoDTensor with Tensor,test=allcases

* Patched python level LoDTensor

* Patched python level LoDTensor

* Merge Tensor into DenseTensor

* Fixed namespace issues,test=allcases

* Fixed merge issues

* Fixed inference issues

* Fixed NPU test issues

* Fixed merge issues

Author: jim19930609

paddle/fluid/distributed/fleet.h

@@ -36,7 +36,6 @@ limitations under the License. */
 
 namespace paddle {
 namespace framework {
-class Tensor;
 class Scope;
 class SelectedRows;
 class Variable;

paddle/fluid/distributed/fleet_executor/dist_model.cc

@@ -145,8 +145,8 @@ bool DistModel::LoadParameters() {
   return true;
 }
 
-void DistModel::Run(const std::vector<framework::Tensor> &input_data,
-                    std::vector<framework::Tensor> *output_data) {
+void DistModel::Run(const std::vector<paddle::framework::Tensor> &input_data,
+                    std::vector<paddle::framework::Tensor> *output_data) {
   /* TODO(fleet exe dev): implement this funct */
 }
 

paddle/fluid/distributed/fleet_executor/dist_model.h

@@ -18,14 +18,14 @@
 #include <vector>
 
 #include "paddle/fluid/distributed/fleet_executor/fleet_executor_desc.pb.h"
+#include "paddle/fluid/framework/tensor.h"
 #include "paddle/fluid/platform/macros.h"
 #include "paddle/fluid/platform/place.h"
 
 namespace paddle {
 namespace framework {
 class ProgramDesc;
 class Scope;
-class Tensor;
 }
 
 namespace distributed {
@@ -45,8 +45,8 @@ class DistModel {
  public:
   explicit DistModel(const DistModelConfig& config) : config_(config) {}
   bool Init();
-  void Run(const std::vector<framework::Tensor>& input_data,
-           std::vector<framework::Tensor>* output_data);
+  void Run(const std::vector<paddle::framework::Tensor>& input_data,
+           std::vector<paddle::framework::Tensor>* output_data);
   ~DistModel() = default;
 
  private:

paddle/fluid/distributed/service/brpc_utils.cc

@@ -20,10 +20,13 @@ limitations under the License. */
 namespace paddle {
 namespace framework {
 class Variable;
-class Tensor;
 }  // namespace framework
 }  // namespace paddle
 
+namespace pten {
+class DenseTensor;
+}  // namespace pten
+
 namespace paddle {
 namespace distributed {
 

paddle/fluid/distributed/test/brpc_service_dense_sgd_test.cc

@@ -31,11 +31,14 @@ class PSClient;
 class PSServer;
 }  // namespace distributed
 namespace framework {
-class Tensor;
 class Variable;
 }  // namespace framework
 }  // namespace paddle
 
+namespace pten {
+class DenseTensor;
+}  // namespace pten
+
 namespace framework = paddle::framework;
 namespace platform = paddle::platform;
 namespace operators = paddle::operators;

paddle/fluid/distributed/test/brpc_service_sparse_sgd_test.cc

@@ -32,11 +32,14 @@ class PSClient;
 class PSServer;
 }  // namespace distributed
 namespace framework {
-class Tensor;
 class Variable;
 }  // namespace framework
 }  // namespace paddle
 
+namespace pten {
+class DenseTensor;
+}  // namespace pten
+
 namespace framework = paddle::framework;
 namespace platform = paddle::platform;
 namespace operators = paddle::operators;

paddle/fluid/eager/legacy/prepared_operator.h

@@ -29,14 +29,17 @@ DECLARE_bool(use_mkldnn);
 
 namespace paddle {
 namespace framework {
-class Tensor;
 class Variable;
 }  // namespace framework
 namespace platform {
 class DeviceContext;
 }  // namespace platform
 }  // namespace paddle
 
+namespace pten {
+class DenseTensor;
+}  // namespace pten
+
 namespace egr {
 namespace legacy {
 

paddle/fluid/framework/copy_same_tensor_test.cc

@@ -68,7 +68,7 @@ static bool CopySameTensorTestMain(const DDim &dims,
     if (sync_copy) {
       TensorCopySync(src_tensor, dst_place, &src_tensor);
     } else {
-      TensorCopy(src_tensor, dst_place, &src_tensor);
+      paddle::framework::TensorCopy(src_tensor, dst_place, &src_tensor);
       platform::DeviceContextPool::Instance().Get(src_place)->Wait();
       platform::DeviceContextPool::Instance().Get(dst_place)->Wait();
     }

paddle/fluid/framework/data_device_transform.cc

@@ -28,8 +28,9 @@ void TransDataDevice(const Tensor &in, const platform::Place &dst_place,
 
   // NOTE(zhiqiu): Special case for CPU->NPU, avoid stream sync.
   if (platform::is_cpu_place(in.place()) && platform::is_npu_place(dst_place)) {
-    TensorCopy(in, dst_place,
-               *platform::DeviceContextPool::Instance().Get(dst_place), out);
+    paddle::framework::TensorCopy(
+        in, dst_place, *platform::DeviceContextPool::Instance().Get(dst_place),
+        out);
     return;
   }
 

paddle/fluid/framework/data_device_transform.h

@@ -21,8 +21,6 @@ limitations under the License. */
 namespace paddle {
 namespace framework {
 
-class Tensor;
-
 void TransDataDevice(const Tensor& in, const platform::Place& dst_place,
                      Tensor* out);