diff --git a/paddle/fluid/operators/interpolate_v2_op.cu b/paddle/fluid/operators/interpolate_v2_op.cu
index d335e1a2f9d58bc4ea95710facf04ad33cd5b4e0..6f8b89ce64523d1cc10be5fbf8e1063d31cca289 100644
--- a/paddle/fluid/operators/interpolate_v2_op.cu
+++ b/paddle/fluid/operators/interpolate_v2_op.cu
@@ -1035,9 +1035,9 @@ static void Interpolate1DCUDAFwd(const framework::ExecutionContext& ctx,
                               : static_cast<float>(new_scale_w);
   }
 
-  int in_cw = c * in_w;
-  int out_cw = c * out_w;
-  int pixelNum = n * out_cw;
+  int64_t in_cw = c * in_w;
+  int64_t out_cw = c * out_w;
+  auto pixelNum = n * out_cw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
@@ -1169,12 +1169,12 @@ static void Interpolate2DCUDAFwd(const framework::ExecutionContext& ctx,
                               : static_cast<float>(new_scale_w);
   }
 
-  int in_hw = in_h * in_w;
-  int out_hw = out_h * out_w;
-  int in_chw = c * in_hw;
-  int out_chw = c * out_hw;
+  int64_t in_hw = in_h * in_w;
+  int64_t out_hw = out_h * out_w;
+  int64_t in_chw = c * in_hw;
+  int64_t out_chw = c * out_hw;
 
-  int pixelNum = n * out_chw;
+  auto pixelNum = n * out_chw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
@@ -1355,12 +1355,12 @@ static void Interpolate3DCUDAFwd(const framework::ExecutionContext& ctx,
                               : static_cast<float>(new_scale_w);
   }
 
-  int in_dhw = in_d * in_h * in_w;
-  int out_dhw = out_d * out_h * out_w;
-  int in_cdhw = c * in_dhw;
-  int out_cdhw = c * out_dhw;
+  int64_t in_dhw = in_d * in_h * in_w;
+  int64_t out_dhw = out_d * out_h * out_w;
+  int64_t in_cdhw = c * in_dhw;
+  int64_t out_cdhw = c * out_dhw;
 
-  int pixelNum = n * out_cdhw;
+  auto pixelNum = n * out_cdhw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
@@ -1456,9 +1456,9 @@ static void Interpolate1DCUDABwd(const framework::ExecutionContext& ctx,
     ratio_w = (align_corners) ? static_cast<float>(in_w - 1) / (out_w - 1)
                               : static_cast<float>(new_scale_w);
   }
-  int in_cw = c * in_w;
-  int out_cw = c * out_w;
-  int pixelNum = n * out_cw;
+  int64_t in_cw = c * in_w;
+  int64_t out_cw = c * out_w;
+  auto pixelNum = n * out_cw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
@@ -1587,11 +1587,11 @@ static void Interpolate2DCUDABwd(const framework::ExecutionContext& ctx,
                               : static_cast<float>(new_scale_w);
   }
 
-  int in_hw = in_h * in_w;
-  int out_hw = out_h * out_w;
-  int in_chw = c * in_hw;
-  int out_chw = c * out_hw;
-  int pixelNum = n * out_chw;
+  int64_t in_hw = in_h * in_w;
+  int64_t out_hw = out_h * out_w;
+  int64_t in_chw = c * in_hw;
+  int64_t out_chw = c * out_hw;
+  auto pixelNum = n * out_chw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
@@ -1773,12 +1773,12 @@ static void Interpolate3DCUDABwd(const framework::ExecutionContext& ctx,
                               : static_cast<float>(new_scale_w);
   }
 
-  int in_dhw = in_d * in_h * in_w;
-  int out_dhw = out_d * out_h * out_w;
-  int in_cdhw = c * in_dhw;
-  int out_cdhw = c * out_dhw;
+  int64_t in_dhw = in_d * in_h * in_w;
+  int64_t out_dhw = out_d * out_h * out_w;
+  int64_t in_cdhw = c * in_dhw;
+  int64_t out_cdhw = c * out_dhw;
 
-  int pixelNum = n * out_cdhw;
+  auto pixelNum = n * out_cdhw;
 
   platform::GpuLaunchConfig config =
       platform::GetGpuLaunchConfig1D(ctx.cuda_device_context(), pixelNum);
diff --git a/python/paddle/nn/functional/common.py b/python/paddle/nn/functional/common.py
index 7cda9a90ff07b0f2306149d5b2d10759091be2d1..323a8af52bccae5dc9f19f13b31c1d2a889e4145 100644
--- a/python/paddle/nn/functional/common.py
+++ b/python/paddle/nn/functional/common.py
@@ -295,13 +295,16 @@ def interpolate(x,
             "align_corners option can only be set with the interpolating modes: linear | bilinear | bicubic | trilinear"
         )
 
-    if resample == 'AREA' and len(x.shape) == 3:
-        return paddle.nn.functional.adaptive_avg_pool1d(x, size)
-
-    if resample == 'AREA' and len(x.shape) == 4:
-        return paddle.nn.functional.adaptive_avg_pool2d(x, size)
-    if resample == 'AREA' and len(x.shape) == 5:
-        return paddle.nn.functional.adaptive_avg_pool3d(x, size)
+    if resample == 'AREA':
+        if isinstance(size, list) or isinstance(size, tuple):
+            if len(size) == 0:
+                raise ValueError("output size can not be empty")
+        if len(x.shape) == 3:
+            return paddle.nn.functional.adaptive_avg_pool1d(x, size)
+        elif len(x.shape) == 4:
+            return paddle.nn.functional.adaptive_avg_pool2d(x, size)
+        elif len(x.shape) == 5:
+            return paddle.nn.functional.adaptive_avg_pool3d(x, size)
 
     helper = LayerHelper('{}_interp_v2'.format(resample_type), **locals())
     dtype = helper.input_dtype(input_param_name='x')
@@ -342,14 +345,13 @@ def interpolate(x,
 
     out_shape = size
     scale = scale_factor
+
     if out_shape is not None and scale is not None:
         raise ValueError("Only one of size or scale_factor should be defined.")
     if out_shape is not None:
-
         if isinstance(out_shape, Variable) and not in_dygraph_mode():
             out_shape.stop_gradient = True
             inputs['OutSize'] = out_shape
-
         else:
             if in_dygraph_mode():
                 if isinstance(out_shape, Variable):