[cherry-pick] Fix numpy2.1 bug in py3.10

paddle/fluid/pybind/tensor.cc

@@ -210,8 +210,16 @@ void BindTensor(pybind11::module &m) {  // NOLINT
   g_framework_tensor_pytype =
       reinterpret_cast<PyTypeObject *>(framework_tensor.ptr());
   framework_tensor
-      .def("__array__",
-           [](phi::DenseTensor &self) { return TensorToPyArray(self); })
+      .def(
+          // TODO(risemeup): Modify the logic of
+          // TensorToPyArray() according to the dtype and copy
+          // parameters.
+          "__array__",
+          [](phi::DenseTensor &self, py::object dtype, py::object copy) {
+            return TensorToPyArray(self);
+          },
+          py::arg("dtype") = py::none(),
+          py::arg("copy") = py::none())
       .def("_ptr",
            [](const phi::DenseTensor &self) {
              return reinterpret_cast<uintptr_t>(self.data());

paddle/fluid/pybind/tensor_py.h

@@ -411,7 +411,7 @@ void SetTensorFromPyArrayT(
                  static_cast<void *>(dst),
                  platform::CPUPlace(),
                  static_cast<const void *>(array.data()),
-                 array.nbytes());
+                 array.size()*sizeof(T));
 #else
     PADDLE_THROW(platform::errors::PermissionDenied(
         "Cannot use XPUPlace in CPU/GPU version, "
@@ -447,7 +447,7 @@ void SetTensorFromPyArrayT(
     phi::DeviceManager::GetDeviceWithPlace(tmp_place)->MemoryCopyH2D(
         reinterpret_cast<void *>(dst),
         const_cast<void *>(reinterpret_cast<const void *>(array.data())),
-        array.nbytes());
+        array.size()*sizeof(T));
     platform::DeviceContextPool &pool = platform::DeviceContextPool::Instance();
     auto &ctx = *pool.Get(place);
     ctx.Wait();
@@ -465,10 +465,10 @@ void SetTensorFromPyArrayT(
       auto dst = self->mutable_data<T>(place);
 #ifdef PADDLE_WITH_HIP
       paddle::platform::GpuMemcpySync(
-          dst, array.data(), array.nbytes(), hipMemcpyHostToDevice);
+          dst, array.data(),array.size()*sizeof(T), hipMemcpyHostToDevice); //NOLINT
 #else
       paddle::platform::GpuMemcpySync(
-          dst, array.data(), array.nbytes(), cudaMemcpyHostToDevice);
+          dst, array.data(), array.size()*sizeof(T), cudaMemcpyHostToDevice);
 #endif
 
     } else if (paddle::platform::is_cuda_pinned_place(place)) {
@@ -617,7 +617,7 @@ void SetUVATensorFromPyArrayImpl(
   cudaHostAlloc(reinterpret_cast<void **>(&data_ptr),
                 need_allocate_size,
                 cudaHostAllocWriteCombined | cudaHostAllocMapped);
-  std::memcpy(data_ptr, array.data(), array.nbytes());
+  std::memcpy(data_ptr, array.data(), array.size()*sizeof(T));
 
   void *cuda_device_pointer = nullptr;
   cudaHostGetDevicePointer(reinterpret_cast<void **>(&cuda_device_pointer),

python/paddle/base/dygraph/tensor_patch_methods.py

@@ -848,7 +848,11 @@ def __nonzero__(self):
     def __bool__(self):
         return self.__nonzero__()
 
-    def __array__(self, dtype=None):
+    def __array__(
+        self,
+        dtype=None,
+        copy=None,
+    ):
         """
         Returns a numpy array shows the value of current Tensor.