[pten] modify reduce_sum reduce_mean args (#38216)

* modify sum mean args

* add GetExpectedPtenKernelArgs for redcue_op

* modify kernel args number

* modify kernel args number

paddle/fluid/operators/reduce_ops/reduce_op.h

@@ -546,6 +546,25 @@ class ReduceOp : public framework::OperatorWithKernel {
     }
     return framework::OpKernelType(input_data_type, ctx.GetPlace());
   }
+
+  framework::KernelSignature GetExpectedPtenKernelArgs(
+      const framework::ExecutionContext& ctx) const override {
+    if (Type() == "reduce_sum") {
+      if (ctx.InputVar("X")->IsType<framework::LoDTensor>()) {
+        return framework::KernelSignature(
+            "sum", {"X"}, {"dim", "keep_dim", "reduce_all", "out_dtype"},
+            {"Out"});
+      }
+    }
+    if (Type() == "reduce_mean") {
+      if (ctx.InputVar("X")->IsType<framework::LoDTensor>()) {
+        return framework::KernelSignature(
+            "mean", {"X"}, {"dim", "keep_dim", "reduce_all"}, {"Out"});
+      }
+    }
+    // TODO(chentianyu03): support other cases after selected rows added
+    return framework::KernelSignature("reduce.unregistered", {}, {}, {});
+  }
 };
 
 class ReduceOpUseInputPlace : public ReduceOp {

paddle/pten/api/include/kernel_signature.h

@@ -71,8 +71,6 @@ using mean_kernel = void (*)(const DeviceContext&,
                              const std::vector<int64_t>&,
                              bool,
                              bool,
-                             DataType,
-                             DataType,
                              DenseTensor*);
 
 using multiply_kernel = void (*)(const DeviceContext&,
@@ -99,7 +97,6 @@ using sum_kernel = void (*)(const DeviceContext&,
                             bool,
                             bool,
                             DataType,
-                            DataType,
                             DenseTensor*);
 
 using subtract_kernel = void (*)(const DeviceContext&,

paddle/pten/include/math.h

@@ -45,9 +45,7 @@ DenseTensor Mean(const ContextT& dev_ctx,
           dev_ctx.GetPlace()),
       std::move(out_meta));
   bool reduce_all = false;
-  DataType out_dtype = pten::DataType::UNDEFINED;
-  Mean<T>(
-      dev_ctx, x, axis, keep_dim, reduce_all, x.dtype(), out_dtype, &dense_out);
+  Mean<T>(dev_ctx, x, axis, keep_dim, reduce_all, &dense_out);
   return dense_out;
 }
 
@@ -57,7 +55,7 @@ DenseTensor Sum(const ContextT& dev_ctx,
                 const std::vector<int64_t>& axis,
                 DataType dtype,
                 bool keep_dim) {
-  auto out_meta = ReduceInferMeta(x.meta(), axis, keep_dim);
+  auto out_meta = ReduceInferMeta(x.meta(), axis, keep_dim, dtype);
   pten::DenseTensor dense_out(
       pten::make_intrusive<paddle::experimental::SharedStorage>(
           dev_ctx.GetPlace()),
@@ -67,12 +65,7 @@ DenseTensor Sum(const ContextT& dev_ctx,
   // so use default value(false) is OK.
   bool reduce_all = false;
 
-  if (x.dtype() == pten::DataType::BOOL || x.dtype() == pten::DataType::INT32 ||
-      x.dtype() == pten::DataType::INT64) {
-    dtype = pten::DataType::INT64;
-  }
-
-  Sum<T>(dev_ctx, x, axis, keep_dim, reduce_all, x.dtype(), dtype, &dense_out);
+  Sum<T>(dev_ctx, x, axis, keep_dim, reduce_all, out_meta.dtype, &dense_out);
   return dense_out;
 }
 

paddle/pten/infermeta/unary.cc

@@ -234,7 +234,8 @@ DenseTensorMeta ReshapeInferMeta(const DenseTensorMeta& x_meta,
 
 DenseTensorMeta ReduceInferMeta(const DenseTensorMeta& x_meta,
                                 const std::vector<int64_t>& axis,
-                                bool keep_dim) {
+                                bool keep_dim,
+                                DataType dtype) {
   bool reduce_all = true;
   std::set<int64_t> dims_set(axis.begin(), axis.end());
   for (int64_t i = 0; i < x_meta.dims.size(); ++i) {
@@ -268,10 +269,16 @@ DenseTensorMeta ReduceInferMeta(const DenseTensorMeta& x_meta,
   }
   DDim out_dim = paddle::framework::make_ddim(out_dim_vector);
 
-  DataType out_dtype = x_meta.dtype;
+  DataType out_dtype;
+  if (dtype != DataType::UNDEFINED) {
+    out_dtype = dtype;
+  } else {
     if (x_meta.dtype == DataType::BOOL || x_meta.dtype == DataType::INT32 ||
         x_meta.dtype == DataType::INT64) {
       out_dtype = DataType::INT64;
+    } else {
+      out_dtype = x_meta.dtype;
+    }
   }
 
   DenseTensorMeta return_meta(out_dtype, out_dim, x_meta.layout);

paddle/pten/infermeta/unary.h

@@ -56,5 +56,6 @@ DenseTensorMeta ReshapeInferMeta(const DenseTensorMeta& x_meta,
 
 DenseTensorMeta ReduceInferMeta(const DenseTensorMeta& x_meta,
                                 const std::vector<int64_t>& axis,
-                                bool keep_dim);
+                                bool keep_dim,
+                                DataType dtype = DataType::UNDEFINED);
 }  // namespace pten

paddle/pten/kernels/cpu/math.cc

@@ -39,9 +39,8 @@ void Mean(const CPUContext& dev_ctx,
           const std::vector<int64_t>& dims,
           bool keep_dim,
           bool reduce_all,
-          DataType in_dtype,
-          DataType out_dtype,
           DenseTensor* out) {
+  auto out_dtype = x.dtype();
   pten::general::Reduce<CPUContext, T, pten::eigen::MeanFunctor>(
       dev_ctx, x, reduce_all, dims, keep_dim, out_dtype, out);
 }
@@ -76,7 +75,6 @@ void Sum(const CPUContext& dev_ctx,
          const std::vector<int64_t>& dims,
          bool keep_dim,
          bool reduce_all,
-         DataType in_dtype,
          DataType out_dtype,
          DenseTensor* out) {
   pten::general::Reduce<CPUContext, T, pten::eigen::SumFunctor>(

paddle/pten/kernels/cpu/math.h

@@ -30,8 +30,6 @@ void Mean(const CPUContext& dev_ctx,
           const std::vector<int64_t>& dims,
           bool keep_dim,
           bool reduce_all,
-          DataType in_dtype,
-          DataType out_dtype,
           DenseTensor* out);
 
 template <typename T>
@@ -67,7 +65,6 @@ void Sum(const CPUContext& dev_ctx,
          const std::vector<int64_t>& dims,
          bool keep_dim,
          bool reduce_all,
-         DataType in_dtype,
          DataType out_dtype,
          DenseTensor* out);
 

paddle/pten/kernels/cuda/math.cu

@@ -68,9 +68,8 @@ void Mean(const CUDAContext& dev_ctx,
           const std::vector<int64_t>& dims,
           bool keep_dim,
           bool reduce_all,
-          DataType in_dtype,
-          DataType out_dtype,
           DenseTensor* out) {
+  auto out_dtype = x.dtype();
   pten::Reduce<T, paddle::operators::CustomMean>(
       dev_ctx, x, reduce_all, dims, keep_dim, out_dtype, out);
 }
@@ -90,7 +89,6 @@ void Sum(const CUDAContext& dev_ctx,
          const std::vector<int64_t>& dims,
          bool keep_dim,
          bool reduce_all,
-         DataType in_dtype,
          DataType out_dtype,
          DenseTensor* out) {
   pten::Reduce<T, paddle::operators::CustomSum>(

paddle/pten/kernels/cuda/math.h

@@ -32,8 +32,6 @@ void Mean(const CUDAContext& dev_ctx,
           const std::vector<int64_t>& dims,
           bool keep_dim,
           bool reduce_all,
-          DataType in_dtype,
-          DataType out_dtype,
           DenseTensor* out);
 
 template <typename T>
@@ -70,7 +68,6 @@ void Sum(const CUDAContext& dev_ctx,
          const std::vector<int64_t>& dims,
          bool keep_dim,
          bool reduce_all,
-         DataType in_dtype,
          DataType out_dtype,
          DenseTensor* out);
 

paddle/pten/tests/api/test_sum_api.cc

@@ -50,7 +50,7 @@ TEST(API, sum) {
   std::vector<int64_t> axis = {0, 1};
 
   // 2. test API
-  auto out = paddle::experimental::sum(x, axis, false);
+  auto out = paddle::experimental::sum(x, axis, DataType::UNDEFINED, false);
   // 3. check result
   ASSERT_EQ(out.dims().size(), 1);
   ASSERT_EQ(out.dims()[0], 1);

python/paddle/utils/code_gen/api.yaml

@@ -79,13 +79,14 @@
     func : matmul
 
 - api : mean
-  args : (const Tensor& x, const std::vector<int64_t>& axis, bool keep_dim)
+  args : (const Tensor& x, const std::vector<int64_t>& axis={}, bool keep_dim=false)
   output : Tensor
   infer_meta : 
     func : ReduceInferMeta
+    param: [x, axis, keep_dim]
   kernel : 
     func : mean
-    param : [x, axis, keep_dim, false, x.dtype(), DataType::UNDEFINED]
+    param : [x, axis, keep_dim, false]
 
 - api : multiply
   args : (const Tensor& x, const Tensor& y)
@@ -130,13 +131,15 @@
     param : [x, y, -1]
 
 - api : sum
-  args : (const Tensor& x, const std::vector<int64_t>& axis, bool keep_dim)
+  args : (const Tensor& x, const std::vector<int64_t>& axis={}, DataType dtype=DataType::UNDEFINED, bool keep_dim=false)
   output : Tensor
   infer_meta : 
     func : ReduceInferMeta
+    param: [x, axis, keep_dim, dtype]
   kernel : 
     func : sum
-    param : [x, axis, keep_dim, false, x.dtype(), DataType::UNDEFINED]
+    param : [x, axis, keep_dim, false, DataType::UNDEFINED]
+    data_type : x
 
 - api : zeros_like
   args : (const Tensor& x, DataType dtype=DataType::UNDEFINED, Backend place=Backend::UNDEFINED, DataLayout layout=DataLayout::UNDEFINED)