precommit

python/mlx/nn/losses.py

@@ -206,7 +206,9 @@ def kl_div_loss(
     return _reduce(loss, reduction)
 
 
-def hinge_loss(predictions: mx.array, targets: mx.array, reduction: str = "none") -> mx.array:
+def hinge_loss(
+    predictions: mx.array, targets: mx.array, reduction: str = "none"
+) -> mx.array:
     """
     Computes the hinge loss between predictions and targets for binary classification tasks.
 
@@ -223,7 +225,12 @@ def hinge_loss(predictions: mx.array, targets: mx.array, reduction: str = "none"
     return _reduce(loss, reduction)
 
 
-def huber_loss(predictions: mx.array, targets: mx.array, delta: float = 1.0, reduction: str = "none") -> mx.array:
+def huber_loss(
+    predictions: mx.array,
+    targets: mx.array,
+    delta: float = 1.0,
+    reduction: str = "none",
+) -> mx.array:
     """
     Computes the Huber loss, a robust loss function for regression tasks.
 
@@ -240,12 +247,14 @@ def huber_loss(predictions: mx.array, targets: mx.array, delta: float = 1.0, red
     error = mx.abs(predictions - targets)
     is_small_error = error < delta
     squared_loss = 0.5 * mx.square(error)
-    linear_loss = delta * error - 0.5 * (delta ** 2)
+    linear_loss = delta * error - 0.5 * (delta**2)
     loss = mx.where(is_small_error, squared_loss, linear_loss)
     return _reduce(loss, reduction)
 
 
-def dice_loss(inputs: mx.array, targets: mx.array, eps: float = 1e-6, reduction: str = "none") -> mx.array:
+def dice_loss(
+    inputs: mx.array, targets: mx.array, eps: float = 1e-6, reduction: str = "none"
+) -> mx.array:
     """
     Computes the Dice loss, useful for binary segmentation tasks.
 
@@ -261,11 +270,18 @@ def dice_loss(inputs: mx.array, targets: mx.array, eps: float = 1e-6, reduction:
     """
     intersection = mx.sum(inputs * targets, axis=1)  # Sum over the feature dimension
     union = mx.sum(inputs, axis=1) + mx.sum(targets, axis=1)
-    dice_score = (2. * intersection + eps) / (union + eps)
+    dice_score = (2.0 * intersection + eps) / (union + eps)
     loss = 1 - dice_score
     return _reduce(loss, reduction)
 
-def focal_loss(inputs: mx.array, targets: mx.array, alpha: float = 0.25, gamma: float = 2.0, reduction: str = "none") -> mx.array:
+
+def focal_loss(
+    inputs: mx.array,
+    targets: mx.array,
+    alpha: float = 0.25,
+    gamma: float = 2.0,
+    reduction: str = "none",
+) -> mx.array:
     """
     Computes the Focal loss, useful for handling class imbalance in binary classification tasks.
 
@@ -286,7 +302,13 @@ def focal_loss(inputs: mx.array, targets: mx.array, alpha: float = 0.25, gamma:
     return _reduce(loss, reduction)
 
 
-def contrastive_loss(embeddings1: mx.array, embeddings2: mx.array, targets: mx.array, margin: float = 1.0, reduction: str = "none") -> mx.array:
+def contrastive_loss(
+    embeddings1: mx.array,
+    embeddings2: mx.array,
+    targets: mx.array,
+    margin: float = 1.0,
+    reduction: str = "none",
+) -> mx.array:
     """
     Computes the Contrastive loss, useful for learning embeddings.
 
@@ -306,7 +328,14 @@ def contrastive_loss(embeddings1: mx.array, embeddings2: mx.array, targets: mx.a
     return _reduce(loss, reduction)
 
 
-def cosine_similarity_loss(embeddings1: mx.array, embeddings2: mx.array, targets: mx.array, eps: float=1e-8, margin: float=0.0, reduction: str = "none") -> mx.array:
+def cosine_similarity_loss(
+    embeddings1: mx.array,
+    embeddings2: mx.array,
+    targets: mx.array,
+    eps: float = 1e-8,
+    margin: float = 0.0,
+    reduction: str = "none",
+) -> mx.array:
     """
     Computes the Cosine Similarity loss, useful for tasks where the angle between embeddings is important.
 
@@ -324,6 +353,10 @@ def cosine_similarity_loss(embeddings1: mx.array, embeddings2: mx.array, targets
     embeddings1_norm = mx.sqrt(mx.sum(mx.square(embeddings1), axis=1) + eps)
     embeddings2_norm = mx.sqrt(mx.sum(mx.square(embeddings2), axis=1) + eps)
 
-    cos_similarity = mx.sum(embeddings1 * embeddings2, axis=1) / (embeddings1_norm * embeddings2_norm)
+    cos_similarity = mx.sum(embeddings1 * embeddings2, axis=1) / (
-    loss = mx.where(targets == 1, 1 - cos_similarity, mx.maximum(0, cos_similarity - margin))
+        embeddings1_norm * embeddings2_norm
+    )
+    loss = mx.where(
+        targets == 1, 1 - cos_similarity, mx.maximum(0, cos_similarity - margin)
+    )
     return _reduce(loss, reduction)