added sigmoid_focal_loss to paddle frontend and refactored mse_loss in paddle frontend

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

@@ -252,16 +252,13 @@ def margin_ranking_loss(input, other, label, margin=0.0, reduction="mean", name=
     return out
 
 
-@with_supported_dtypes({"2.4.2 and below": ("float32", "float64")}, "paddle")
+@with_supported_dtypes({"2.5.1 and below": ("float32", "float64")}, "paddle")
 @inputs_to_ivy_arrays
 def mse_loss(input, label, reduction="mean", name=None):
     reduction = _get_reduction_func(reduction)
     ret = ivy.square(input - label)
     ret = reduction(ret)
 
-    if ret.shape == ():
-        ret = ret.expand_dims()
-
     return paddle.to_tensor(ret)
 
 
@@ -298,6 +295,52 @@ def nll_loss(
     return output
 
 
+@with_supported_dtypes({"2.5.1 and below": ("float32", "float64")}, "paddle")
+@to_ivy_arrays_and_back
+def sigmoid_focal_loss(
+    logit,
+    label,
+    normalizer=None,
+    alpha=0.25,
+    gamma=2.0,
+    reduction="sum",
+    name=None,
+):
+    if reduction not in ["sum", "mean", "none"]:
+        raise ValueError(
+            "The value of 'reduction' in sigmoid_focal_loss should be 'sum', 'mean' or"
+            f" 'none', but received {reduction}, which is not allowed."
+        )
+
+    if normalizer is not None and normalizer.ndim > 1:
+        raise ValueError(
+            "Expected zero or one dimension of normalizer in sigmoid_focal_loss but"
+            f" got {normalizer.ndim}."
+        )
+
+    if not isinstance(logit, ivy.Array):
+        logit = ivy.array(logit)
+
+    if not isinstance(label, ivy.Array):
+        label = ivy.array(label)
+
+    pred = ivy.sigmoid(logit)
+    loss = -(
+        label * alpha * ivy.pow((1 - pred), gamma) * ivy.log(pred)
+        + (1 - label) * (1 - alpha) * ivy.pow(pred, gamma) * ivy.log(1 - pred)
+    )
+
+    if normalizer is not None:
+        loss /= normalizer
+
+    if reduction == "sum":
+        return ivy.sum(loss)
+    elif reduction == "mean":
+        return ivy.mean(loss)
+
+    return loss
+
+
 @with_supported_dtypes({"2.5.1 and below": ("float32", "float64")}, "paddle")
 @to_ivy_arrays_and_back
 def smooth_l1_loss(

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

@@ -468,6 +468,81 @@ def test_paddle_nll_loss(
     )
 
 
+@handle_frontend_test(
+    fn_tree="paddle.nn.functional.sigmoid_focal_loss",
+    dtype_and_x=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("float"),
+        num_arrays=1,
+        shared_dtype=False,
+        min_num_dims=1,
+        min_dim_size=1,
+    ),
+    dtype_and_normalizer=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("float"),
+        num_arrays=1,
+        shared_dtype=True,
+        min_num_dims=1,
+        min_dim_size=1,
+        max_num_dims=1,
+        max_dim_size=1,
+    ),
+    dtype_and_labels=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("float"),
+        num_arrays=1,
+        shared_dtype=False,
+        min_num_dims=1,
+        min_dim_size=1,
+        min_value=0,
+        max_value=1,
+    ),
+    alpha=st.floats(
+        min_value=0.0,
+        max_value=1.0,
+    ),
+    gamma=st.floats(
+        min_value=0.0,
+        max_value=5.0,
+    ),
+    reduction=st.sampled_from(["mean", "sum", "none"]),
+)
+def test_paddle_sigmoid_focal_loss(
+    dtype_and_x,
+    dtype_and_normalizer,
+    dtype_and_labels,
+    alpha,
+    gamma,
+    reduction,
+    on_device,
+    fn_tree,
+    frontend,
+    test_flags,
+    backend_fw,
+):
+    x_dtype, x = dtype_and_x
+    normalizer_dtype, normalizer = dtype_and_normalizer
+    label_dtype, labels = dtype_and_labels
+    normalizer = [norm.reshape(-1) for norm in normalizer]
+    labels = ivy.array(labels, dtype=ivy.int64)
+    helpers.test_frontend_function(
+        input_dtypes=[ivy.int64]
+        + [ivy.float64]
+        + x_dtype
+        + normalizer_dtype
+        + label_dtype,
+        backend_to_test=backend_fw,
+        frontend=frontend,
+        test_flags=test_flags,
+        fn_tree=fn_tree,
+        on_device=on_device,
+        logit=x[0],
+        label=labels[0],
+        alpha=alpha,
+        gamma=gamma,
+        normalizer=normalizer[0],
+        reduction=reduction,
+    )
+
+
 # smooth_l1_loss
 @handle_frontend_test(
     fn_tree="paddle.nn.functional.smooth_l1_loss",