【Hackathon No.48】为 Paddle meshgrid 算子实现 float16 数据类型支持 (#53284)

paddle/phi/kernels/gpu/meshgrid_grad_kernel.cu.cc

@@ -22,6 +22,7 @@ PD_REGISTER_KERNEL(meshgrid_grad,
                    GPU,
                    ALL_LAYOUT,
                    phi::MeshgridGradKernel,
+                   phi::dtype::float16,
                    float,
                    double,
                    int,

paddle/phi/kernels/gpu/meshgrid_kernel.cu.cc

@@ -22,6 +22,7 @@ PD_REGISTER_KERNEL(meshgrid,
                    GPU,
                    ALL_LAYOUT,
                    phi::MeshgridKernel,
+                   phi::dtype::float16,
                    float,
                    double,
                    int,

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

@@ -76,6 +76,14 @@ class TestMeshgridOp2(TestMeshgridOp):
         return [100, 300]
 
 
+class TestMeshgridOp2Fp16(TestMeshgridOp):
+    def get_x_shape(self):
+        return [100, 300]
+
+    def get_dtype(self):
+        return np.float16
+
+
 class TestMeshgridOp3(unittest.TestCase):
     def test_api(self):
         x = paddle.static.data(shape=[100], dtype='int32', name='x')

python/paddle/tensor/creation.py

@@ -1512,7 +1512,7 @@ def meshgrid(*args, **kwargs):
 
     Args:
         *args(Tensor|list of Tensor) : tensors (tuple(list) of tensor): the shapes of input k tensors are (N1,),
-            (N2,),..., (Nk,). Support data types: ``float64``, ``float32``, ``int32``, ``int64``.
+            (N2,),..., (Nk,). Support data types: ``float64``, ``float16``, ``float32``, ``int32``, ``int64``.
         **kwargs (optional): Currently, only accept name in **kwargs
             The default value is None. Normally there is no need for
             user to set this property. For more information, please refer to :ref:`api_guide_Name`.