Fix opitimize optimize (#69145)

Author: co63oc

paddle/fluid/pybind/compiled_program.cc

@@ -770,7 +770,7 @@ void BindCompiledProgram(pybind11::module &m) {  // NOLINT
                   "or True"));
             }
           },
-          R"DOC((bool, optional): memory opitimize aims to save total memory
+          R"DOC((bool, optional): memory optimize aims to save total memory
                 consumption, set to True to enable it.
 
                 Default None. None means framework would choose to use or not use

paddle/fluid/pybind/eager.cc

@@ -764,7 +764,7 @@ PyDoc_STRVAR(  // NOLINT
 
 Tensor is the basic data structure in PaddlePaddle. There are some ways to create a Tensor:
 
-- Use the exsiting ``data`` to create a Tensor, please refer to :ref:`api_paddle_to_tensor`.
+- Use the existing ``data`` to create a Tensor, please refer to :ref:`api_paddle_to_tensor`.
 - Create a Tensor with a specified ``shape``, please refer to :ref:`api_paddle_ones`,
   :ref:`api_paddle_zeros`, :ref:`api_paddle_full`.
 - Create a Tensor with the same ``shape`` and ``dtype`` as other Tensor, please refer to

paddle/fluid/pybind/eager_generator.cc

@@ -671,7 +671,7 @@ static void PurifyForwardOpProto(const proto::OpProto& op_proto,
     }
   }
 
-  /* ------ Maping forward slot name to fwd position ------ */
+  /* ------ Mapping forward slot name to fwd position ------ */
   size_t in_pos = 0;
   for (const auto& var : *in_vars) {
     VLOG(6) << "Mapping input tensor: " << var.name()
@@ -1877,7 +1877,7 @@ static std::pair<std::string, std::string> GenerateForwardFunctionContents(
   trace_op_body_str += trace_op_str;
   trace_op_body_str += "\n";
 
-  // [Generation] Log memory infomation
+  // [Generation] Log memory information
   const char* LOG_MEMORY_INFO_TEMPLATE =
       " // Log memory information\n"
       "  "

paddle/fluid/pybind/eager_method.cc

@@ -719,7 +719,7 @@ PyDoc_STRVAR(tensor_method_clone__doc__,  // NOLINT
 
 Returns a new Tensor, which is clone of origin Tensor, and it remains in the current graph.
 It will always have a Tensor copy.
-Tn addition, the cloned Tensor provides gradient propagation.
+In addition, the cloned Tensor provides gradient propagation.
 
 Returns:
     Tensor, The cloned Tensor.
@@ -841,7 +841,7 @@ PyDoc_STRVAR(tensor_clear_gradient__doc__,  // NOLINT
 --
 
 Only for Tensor that has gradient, normally we use this for Parameters since
-other temporary Tensor doesen't has gradient.
+other temporary Tensor doesn't has gradient.
 
 The Gradient of current Tensor will be set to ``0`` elementwise or ``None``.
 

paddle/fluid/pybind/pybind.cc

@@ -936,7 +936,7 @@ void BindVjp(pybind11::module *m) {
             PADDLE_ENFORCE_EQ(inputs[idx].size(),
                               vjp_res[grad_index].size(),
                               common::errors::InvalidArgument(
-                                  "The size of inouts[%d] should be the "
+                                  "The size of inputs[%d] should be the "
                                   "same as vjp_res[%d] size.",
                                   idx,
                                   grad_index));
@@ -1385,7 +1385,7 @@ PYBIND11_MODULE(libpaddle, m) {
            Return the registered kernels in paddle.
 
            Args:
-               lib[string]: the libarary, could be 'phi', 'fluid' and 'all'.
+               lib[string]: the library, could be 'phi', 'fluid' and 'all'.
            )DOC");
 
   m.def(
@@ -1437,7 +1437,7 @@ PYBIND11_MODULE(libpaddle, m) {
            Return the registered kernels in phi.
 
            Args:
-               kernel_registered_type[string]: the libarary, could be 'function', 'structure', and 'all'.
+               kernel_registered_type[string]: the library, could be 'function', 'structure', and 'all'.
            )DOC");
 
   // NOTE(Aganlengzi): KernelFactory static instance is initialized BEFORE
@@ -1825,7 +1825,7 @@ All parameter, weight, gradient are variables in Paddle.
           VLOG(3) << "need skip: " << need_skip << std::endl;
           if (paddle::prim::PrimCommonUtils::IsBwdPrimEnabled()) {
             if ((grad_comp_op_maker != nullptr) && (!need_skip)) {
-              VLOG(3) << "Prim Flag Open: Runing composite grad fun for "
+              VLOG(3) << "Prim Flag Open: Running composite grad fun for "
                       << op_desc.Type();
               grad_op_descs = grad_comp_op_maker(op_desc,
                                                  no_grad_set,
@@ -1838,12 +1838,12 @@ All parameter, weight, gradient are variables in Paddle.
             }
           } else {
             if (grad_op_maker != nullptr) {
-              VLOG(6) << "Prim Flag Close: Runing origin grad fun for "
+              VLOG(6) << "Prim Flag Close: Running origin grad fun for "
                       << op_desc.Type();
               grad_op_descs = grad_op_maker(
                   op_desc, no_grad_set, &grad_to_var, grad_sub_block);
             } else {
-              VLOG(6) << "Prim Flag Close: Runing composite grad fun for "
+              VLOG(6) << "Prim Flag Close: Running composite grad fun for "
                       << op_desc.Type();
               grad_op_descs = grad_comp_op_maker(op_desc,
                                                  no_grad_set,

paddle/fluid/pybind/tensor.cc

@@ -700,7 +700,7 @@ void BindTensor(pybind11::module &m) {  // NOLINT
              return dst;
            })
       .def("_copy", [](const phi::DenseTensor &self, const phi::Place &place) {
-        // follow fetch_op's inplementation
+        // follow fetch_op's implementation
         phi::DenseTensor dst;
         if (self.IsInitialized() && self.numel() > 0) {
           TensorCopySync(self, place, &dst);
@@ -759,7 +759,7 @@ void BindTensor(pybind11::module &m) {  // NOLINT
 
            Params:
                tensor: Shared Cuda IPC tensor.
-               tuple: contrains data size, data type,
+               tuple: contains data size, data type,
                       tensor dims, lod information, device index.
 
        )DOC")
@@ -812,7 +812,7 @@ void BindTensor(pybind11::module &m) {  // NOLINT
            Serialize GPU Tensor by cudaIpcMemHandle.
 
            Returns:
-               tuple: contrains handle, data size, data type,
+               tuple: contains handle, data size, data type,
                       tensor dims, lod information, device index.
 
            Examples:
@@ -858,7 +858,7 @@ void BindTensor(pybind11::module &m) {  // NOLINT
            Deserialize GPU lod tensor from cudaIpcMemHandle.
 
            Params:
-               tuple: contrains handle, data size, data type,
+               tuple: contains handle, data size, data type,
                       tensor dims, lod information, device index.
 
            Examples:
@@ -942,7 +942,7 @@ void BindTensor(pybind11::module &m) {  // NOLINT
            If the tensor is not in shared memory, we will copy it first.
 
            Returns:
-               tuple: contrains ipc name, data size, data type,
+               tuple: contains ipc name, data size, data type,
                       tensor dims and lod imformation.
 
            Examples:

python/paddle/tensor/linalg.py

@@ -374,7 +374,7 @@ def __check_input(x, y):
                     dtype='bfloat16'
                 )
             else:
-                raise ValueError("The output_dtype must be float16 or bfloa16")
+                raise ValueError("The output_dtype must be float16 or bfloat16")
 
             helper.append_op(
                 type='fp8_fp8_half_gemm_fused',
@@ -408,7 +408,7 @@ def __check_input(x, y):
                     bias, 'bias', ['bfloat16'], 'fp8_fp8_half_gemm_fused'
                 )
             else:
-                raise ValueError("The output_dtype must be float16 or bfloa16")
+                raise ValueError("The output_dtype must be float16 or bfloat16")
 
             helper = LayerHelper('fp8_fp8_half_gemm_fused', **locals())
 
@@ -419,7 +419,7 @@ def __check_input(x, y):
                     dtype='bfloat16'
                 )
             else:
-                raise ValueError("The output_dtype must be float16 or bfloa16")
+                raise ValueError("The output_dtype must be float16 or bfloat16")
 
             helper.append_op(
                 type='fp8_fp8_half_gemm_fused',