reorder MatrixRank (#52925)

* reorder MatrixRank

* fix

* fix

* fix

* fix

* fix

paddle/phi/api/yaml/legacy_ops.yaml

@@ -755,11 +755,11 @@
   backward : matmul_grad
 
 - op : matrix_rank
-  args : (Tensor x, float tol, bool hermitian=false, bool use_default_tol=true)
+  args : (Tensor x, float tol, bool use_default_tol=true, bool hermitian=false)
   output : Tensor(out)
   infer_meta :
     func : MatrixRankInferMeta
-    param : [x, hermitian, use_default_tol]
+    param : [x, use_default_tol, hermitian]
   kernel :
     func : matrix_rank
 

paddle/phi/api/yaml/static_ops.yaml

@@ -222,7 +222,7 @@
   output : Tensor(out)
   infer_meta :
     func : MatrixRankStaticInferMeta
-    param : [x, tol_tensor, hermitian, use_default_tol]
+    param : [x, tol_tensor, use_default_tol, hermitian]
   optional : tol_tensor
   kernel :
     func : matrix_rank {dense -> dense},

paddle/phi/infermeta/binary.cc

@@ -2207,7 +2207,7 @@ void MatrixRankStaticInferMeta(const MetaTensor& x,
   if (atol_tensor) {
     MatrixRankTolInferMeta(x, atol_tensor, use_default_tol, hermitian, out);
   } else {
-    MatrixRankInferMeta(x, hermitian, use_default_tol, out);
+    MatrixRankInferMeta(x, use_default_tol, hermitian, out);
   }
 }
 

paddle/phi/infermeta/unary.cc

@@ -2006,8 +2006,8 @@ void LUInferMeta(const MetaTensor& x,
 }
 
 void MatrixRankInferMeta(const MetaTensor& x,
-                         bool hermitian,
                          bool use_default_tol,
+                         bool hermitian,
                          MetaTensor* out) {
   auto dim_x = x.dims();
   PADDLE_ENFORCE_GE(dim_x.size(),

paddle/phi/infermeta/unary.h

@@ -272,8 +272,8 @@ void LUInferMeta(const MetaTensor& x,
 void MatrixPowerInferMeta(const MetaTensor& x, int n, MetaTensor* out);
 
 void MatrixRankInferMeta(const MetaTensor& x,
-                         bool hermitian,
                          bool use_default_tol,
+                         bool hermitian,
                          MetaTensor* out);
 
 void MaxOutInferMeta(const MetaTensor& x,

paddle/phi/kernels/cpu/matrix_rank_kernel.cc

@@ -24,8 +24,8 @@ template <typename T, typename Context>
 void MatrixRankKernel(const Context& dev_ctx,
                       const DenseTensor& x,
                       float tol,
-                      bool hermitian,
                       bool use_default_tol,
+                      bool hermitian,
                       DenseTensor* out) {
   DenseTensor atol_tensor;
   if (use_default_tol) {

paddle/phi/kernels/gpu/matrix_rank_kernel.cu

@@ -27,8 +27,8 @@ template <typename T, typename Context>
 void MatrixRankKernel(const Context& dev_ctx,
                       const DenseTensor& x,
                       float tol,
-                      bool hermitian,
                       bool use_default_tol,
+                      bool hermitian,
                       DenseTensor* out) {
   DenseTensor atol_tensor;
   if (use_default_tol) {

paddle/phi/kernels/matrix_rank_kernel.h

@@ -22,8 +22,8 @@ template <typename T, typename Context>
 void MatrixRankKernel(const Context& dev_ctx,
                       const DenseTensor& x,
                       float tol,
-                      bool hermitian,
                       bool use_default_tol,
+                      bool hermitian,
                       DenseTensor* out);
 
 }  // namespace phi

paddle/phi/ops/compat/matrix_rank_sig.cc

@@ -34,8 +34,8 @@ KernelSignature MatrixRankOpArgumentMapping(const ArgumentMappingContext& ctx) {
                            {"X"},
                            {
                                "tol",
-                               "hermitian",
                                "use_default_tol",
+                               "hermitian",
                            },
                            {"Out"});
   }

python/paddle/fluid/tests/unittests/test_matrix_rank_op.py

@@ -26,7 +26,7 @@ SEED = 2049
 np.random.seed(SEED)
 
 
-def matrix_rank_wraper(x, tol=None, hermitian=False, use_default_tol=True):
+def matrix_rank_wraper(x, tol=None, use_default_tol=True, hermitian=False):
     return paddle.linalg.matrix_rank(x, tol, hermitian)
 
 

python/paddle/tensor/linalg.py

@@ -1505,7 +1505,7 @@ def matrix_rank(x, tol=None, hermitian=False, name=None):
         else:
             tol_attr = float(tol)
             use_default_tol = False
-        return _C_ops.matrix_rank(x, tol_attr, hermitian, use_default_tol)
+        return _C_ops.matrix_rank(x, tol_attr, use_default_tol, hermitian)
     else:
         inputs = {}
         attrs = {}