[0-size Tensor No.138] Add 0-size Tensor support for mv

paddle/phi/kernels/cpu/mv_grad_kernel.cc

@@ -16,6 +16,7 @@
 
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/funcs/blas/blas.h"
 
 namespace phi {
@@ -30,6 +31,21 @@ void MvGradKernel(const Context& dev_ctx,
   auto dout = out_grad;
   auto dx = x_grad;
   auto dvec = vec_grad;
+  if (x.numel() == 0 || vec.numel() == 0) {
+    if (dx) {
+      phi::Full<T, Context>(dev_ctx,
+                            phi::IntArray(common::vectorize(dx->dims())),
+                            static_cast<T>(0),
+                            dx);
+    }
+    if (dvec) {
+      phi::Full<T, Context>(dev_ctx,
+                            phi::IntArray(common::vectorize(dvec->dims())),
+                            static_cast<T>(0),
+                            dvec);
+    }
+    return;
+  }
 
   const auto& dim_x = x.dims();
   int m = static_cast<int>(dim_x[0]);

paddle/phi/kernels/gpu/mv_grad_kernel.cu

@@ -16,6 +16,7 @@
 
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
+#include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/funcs/blas/blas.h"
 
 namespace phi {
@@ -41,6 +42,21 @@ void MvGradKernel(const Context &dev_ctx,
   auto dout = out_grad;
   auto dx = x_grad;
   auto dvec = vec_grad;
+  if (x.numel() == 0 || vec.numel() == 0) {
+    if (dx) {
+      phi::Full<T, Context>(dev_ctx,
+                            phi::IntArray(common::vectorize(dx->dims())),
+                            static_cast<T>(0),
+                            dx);
+    }
+    if (dvec) {
+      phi::Full<T, Context>(dev_ctx,
+                            phi::IntArray(common::vectorize(dvec->dims())),
+                            static_cast<T>(0),
+                            dvec);
+    }
+    return;
+  }
 
   auto dim_x = x.dims();
   int m = dim_x[0];

paddle/phi/kernels/impl/mv_kernel_impl.h

@@ -14,8 +14,8 @@
 
 #pragma once
 
+#include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/funcs/blas/blas.h"
-
 namespace phi {
 
 template <typename T, typename Context>
@@ -29,6 +29,17 @@ void MvKernel(const Context& dev_ctx,
   const T* x_data = x.data<T>();
   const T* vec_data = vec.data<T>();
   T* out_data = dev_ctx.template Alloc<T>(out);
+  if (out && out->numel() == 0) {
+    return;
+  }
+  // x.shape [10, 0], vec.shape [0], out.shape [10]
+  if (vec.numel() == 0) {
+    phi::Full<T, Context>(dev_ctx,
+                          phi::IntArray(common::vectorize(out->dims())),
+                          static_cast<T>(0),
+                          out);
+    return;
+  }
 
   auto blas = phi::funcs::GetBlas<Context, T>(dev_ctx);
 

test/legacy_test/test_mv_op.py

@@ -39,6 +39,24 @@ def init_config(self):
         self.vec = np.random.random(100).astype("float64")
 
 
+class TestMVOp_ZeroSize1(TestMVOp):
+    def init_config(self):
+        self.x = np.random.random((0, 100)).astype("float64")
+        self.vec = np.random.random(100).astype("float64")
+
+
+class TestMVOp_ZeroSize2(TestMVOp):
+    def init_config(self):
+        self.x = np.random.random((100, 0)).astype("float64")
+        self.vec = np.random.random(0).astype("float64")
+
+
+class TestMVOp_ZeroSize3(TestMVOp):
+    def init_config(self):
+        self.x = np.random.random((0, 0)).astype("float64")
+        self.vec = np.random.random(0).astype("float64")
+
+
 class TestMVAPI(unittest.TestCase):
     def test_dygraph_api_out(self):
         paddle.disable_static()