feat: created func for bilinear

ivy/functional/frontends/paddle/nn/functional/common.py

@@ -4,6 +4,14 @@
 from ivy.functional.frontends.paddle.func_wrapper import to_ivy_arrays_and_back
 
 
+@to_ivy_arrays_and_back
+@with_supported_dtypes({"2.6.0 and below": ("float32", "float64")}, "paddle")
+def bilinear(x1, x2, weight, bias=None, name=None):
+    x2_transposed = ivy.swapaxes(x2, -1, -2)
+    result = ivy.linear(ivy.multiply(x1, x2_transposed), weight, bias=bias)
+    return x1, x2_transposed, weight, bias, result
+
+
 @to_ivy_arrays_and_back
 @with_supported_dtypes({"2.6.0 and below": ("float32", "float64")}, "paddle")
 def cosine_similarity(x1, x2, *, axis=1, eps=1e-08):

ivy_tests/test_ivy/test_frontends/test_paddle/test_nn/test_functional/test_common.py

@@ -217,6 +217,41 @@ def paddle_unfold_handler(draw, dtype):
 # ------------ #
 
 
+# test function for bilinear
+@handle_frontend_test(
+    fn_tree="paddle.nn.functional.common.bilinear",
+    dtype_x1_x2_weight_bias=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("valid", full=False),
+        num_arrays=5,
+    ),
+)
+def test_paddle_bilinear(
+    *,
+    dtype_x1_x2_weight_bias,
+    on_device,
+    fn_tree,
+    backend_fw,
+    frontend,
+    test_flags,
+):
+    dtype, x1, x2, weight, bias = dtype_x1_x2_weight_bias
+    x2_transposed = ivy.swapaxes(x2, -1, -2)
+    result = ivy.linear(ivy.multiply(x1, x2_transposed), weight, bias=bias)
+    helpers.test_frontend_function(
+        input_dtypes=dtype,
+        frontend=frontend,
+        backend_to_test=backend_fw,
+        test_flags=test_flags,
+        fn_tree=fn_tree,
+        on_device=on_device,
+        x1=x1,
+        x2=x2,
+        weight=weight,
+        bias=bias,
+        result=result,
+    )
+
+
 # Cosine Similarity
 @handle_frontend_test(
     fn_tree="paddle.nn.functional.common.cosine_similarity",