set src_idx > 0 for bilinear_interp_op (#18238)

* set src_idx > 0, test=develop

* add unittest and cu, test=develop

paddle/fluid/operators/interpolate_op.cu

@@ -108,8 +108,10 @@ __global__ void KeBilinearInterpFw(
                          : static_cast<int>(ratio_h * out_img_idy);
     in_img_idy = (in_img_idy > 0) ? in_img_idy : 0;
     int h_id = (in_img_idy < in_img_h - 1) ? 1 : 0;
-    T h1lambda = align_flag ? ratio_h * (out_img_idy + 0.5) - 0.5 - in_img_idy
-                            : ratio_h * out_img_idy - in_img_idy;
+    T src_h = ratio_h * (out_img_idy + 0.5) - 0.5;
+    src_h = (src_h > 0) ? src_h : 0;
+    T h1lambda =
+        align_flag ? src_h - in_img_idy : ratio_h * out_img_idy - in_img_idy;
     T h2lambda = 1.f - h1lambda;
 
     int out_img_idx = tid % out_img_w;
@@ -118,8 +120,10 @@ __global__ void KeBilinearInterpFw(
                          : static_cast<int>(ratio_w * out_img_idx);
     in_img_idx = (in_img_idx > 0) ? in_img_idx : 0;
     int w_id = (in_img_idx < in_img_w - 1) ? 1 : 0;
-    T w1lambda = align_flag ? ratio_w * (out_img_idx + 0.5) - 0.5 - in_img_idx
-                            : ratio_w * out_img_idx - in_img_idx;
+    T src_w = ratio_w * (out_img_idx + 0.5) - 0.5;
+    src_w = (src_w > 0) ? src_w : 0;
+    T w1lambda =
+        align_flag ? src_w - in_img_idx : ratio_w * out_img_idx - in_img_idx;
     T w2lambda = 1.f - w1lambda;
 
     const T* in_pos = &in[out_id_h * input_w + channel_id * in_img_size +
@@ -156,8 +160,10 @@ __global__ void KeBilinearInterpBw(
                                 : ratio_h * out_img_idy;
     in_img_idy = (in_img_idy > 0) ? in_img_idy : 0;
     int h_id = (in_img_idy < in_img_h - 1) ? 1 : 0;
-    T h1lambda = align_flag ? ratio_h * (out_img_idy + 0.5) - 0.5 - in_img_idy
-                            : ratio_h * out_img_idy - in_img_idy;
+    T src_h = ratio_h * (out_img_idy + 0.5) - 0.5;
+    src_h = (src_h > 0) ? src_h : 0;
+    T h1lambda =
+        align_flag ? src_h - in_img_idy : ratio_h * out_img_idy - in_img_idy;
 
     T h2lambda = 1.f - h1lambda;
 
@@ -166,8 +172,10 @@ __global__ void KeBilinearInterpBw(
                                 : ratio_w * out_img_idx;
     in_img_idx = (in_img_idx > 0) ? in_img_idx : 0;
     int w_id = (in_img_idx < in_img_w - 1) ? 1 : 0;
-    T w1lambda = align_flag ? ratio_w * (out_img_idx + 0.5) - 0.5 - in_img_idx
-                            : ratio_w * out_img_idx - in_img_idx;
+    T src_w = ratio_w * (out_img_idx + 0.5) - 0.5;
+    src_w = (src_w > 0) ? src_w : 0;
+    T w1lambda =
+        align_flag ? src_w - in_img_idx : ratio_w * out_img_idx - in_img_idx;
     T w2lambda = 1.f - w1lambda;
 
     T* in_pos = &in[out_id_h * input_w + channel_id * in_img_size +

paddle/fluid/operators/interpolate_op.h

@@ -73,8 +73,9 @@ static void BilinearInterpolation(const Tensor& input, Tensor* output,
                          : static_cast<int>(ratio_h * k);
     y_n = (y_n > 0) ? y_n : 0;
     int y_s = (y_n + 1) < (in_h - 1) ? (y_n + 1) : (in_h - 1);
-    float d_n =
-        align_flag ? ratio_h * (k + 0.5) - 0.5 - y_n : ratio_h * k - y_n;
+    float idx_src_y = ratio_h * (k + 0.5) - 0.5;
+    idx_src_y = (idx_src_y > 0) ? idx_src_y : 0;
+    float d_n = align_flag ? idx_src_y - y_n : ratio_h * k - y_n;
     float d_s = 1.f - d_n;
     {
       vy_n[k] = y_n;
@@ -99,8 +100,9 @@ static void BilinearInterpolation(const Tensor& input, Tensor* output,
                   : static_cast<int>(ratio_w * l);
     x_w = (x_w > 0) ? x_w : 0;
     int x_e = (x_w + 1) < (in_w - 1) ? (x_w + 1) : (in_w - 1);
-    float d_w =
-        align_flag ? ratio_w * (l + 0.5) - 0.5 - x_w : ratio_w * l - x_w;
+    float idx_src_x = ratio_w * (l + 0.5) - 0.5;
+    idx_src_x = (idx_src_x > 0) ? idx_src_x : 0;
+    float d_w = align_flag ? idx_src_x - x_w : ratio_w * l - x_w;
     float d_e = 1.f - d_w;
     {
       vx_w[l] = x_w;
@@ -170,8 +172,9 @@ static void BilinearInterpolationGrad(const Tensor& output_grad,
                          : static_cast<int>(ratio_h * k);
     y_n = (y_n > 0) ? y_n : 0;
     int y_s = (y_n + 1) < (in_h - 1) ? (y_n + 1) : (in_h - 1);
-    float d_n =
-        align_flag ? ratio_h * (k + 0.5) - 0.5 - y_n : ratio_h * k - y_n;
+    float idx_src_y = ratio_h * (k + 0.5) - 0.5;
+    idx_src_y = (idx_src_y > 0) ? idx_src_y : 0;
+    float d_n = align_flag ? idx_src_y - y_n : ratio_h * k - y_n;
     float d_s = 1.f - d_n;
 
     for (int l = 0; l < out_w; l++) {
@@ -179,8 +182,9 @@ static void BilinearInterpolationGrad(const Tensor& output_grad,
                            : static_cast<int>(ratio_w * l);
       x_w = (x_w > 0) ? x_w : 0;
       int x_e = (x_w + 1) < (in_w - 1) ? (x_w + 1) : (in_w - 1);
-      float d_w =
-          align_flag ? ratio_w * (l + 0.5) - 0.5 - x_w : ratio_w * l - x_w;
+      float idx_src_x = ratio_w * (l + 0.5) - 0.5;
+      idx_src_x = (idx_src_x > 0) ? idx_src_x : 0;
+      float d_w = align_flag ? idx_src_x - x_w : ratio_w * l - x_w;
       float d_e = 1.f - d_w;
 
       for (int i = 0; i < n; i++) {    // loop for batches

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

@@ -59,7 +59,8 @@ def bilinear_interp_np(input,
         h = max(0, h)
         hid = 1 if h < in_h - 1 else 0
         if (align_mode == 0 and not align_corners):
-            h1lambda = ratio_h * (i + 0.5) - 0.5 - h
+            idx_src_h = max(ratio_h * (i + 0.5) - 0.5, 0)
+            h1lambda = idx_src_h - h
         else:
             h1lambda = ratio_h * i - h
         h2lambda = 1.0 - h1lambda
@@ -71,7 +72,8 @@ def bilinear_interp_np(input,
             w = max(0, w)
             wid = 1 if w < in_w - 1 else 0
             if (align_mode == 0 and not align_corners):
-                w1lambda = ratio_w * (j + 0.5) - 0.5 - w
+                idx_src_w = max(ratio_w * (j + 0.5) - 0.5, 0)
+                w1lambda = idx_src_w - w
             else:
                 w1lambda = ratio_w * j - w
             w2lambda = 1.0 - w1lambda
@@ -335,5 +337,16 @@ class TestBilinearInterpScale3(TestBilinearInterpOp):
         self.align_mode = 1
 
 
+class TestBilinearInterpZero(TestBilinearInterpOp):
+    def init_test_case(self):
+        self.interp_method = 'bilinear'
+        self.input_shape = [2, 3, 5, 7]
+        self.out_h = 60
+        self.out_w = 25
+        self.scale = 0.2
+        self.align_corners = False
+        self.align_mode = 0
+
+
 if __name__ == "__main__":
     unittest.main()