Formats codebase with rustfmt

Applies rustfmt configuration to ensure consistent code style, including import organization, line breaks for readability, and trailing newlines. Improves maintainability and adheres to Rust formatting standards.

Author: nusu-github

crates/birefnet-backbones/src/lib.rs

@@ -277,4 +277,4 @@ mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}

crates/birefnet-extra-ops/src/lib.rs

@@ -23,4 +23,4 @@ mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}

crates/birefnet-inference/src/lib.rs

@@ -11,4 +11,4 @@ mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}

crates/birefnet-loss/src/birefnet_loss.rs

@@ -246,15 +246,15 @@ impl<B: Backend> BiRefNetLoss<B> {
         let mut total_loss = pixel_loss;
 
         // Add classification loss if enabled and predictions are provided
-        if let Some(ref cls_loss) = self.classification_loss {
-            if let (Some(class_preds), Some(class_targets)) = (class_preds, class_targets) {
-                let cls_loss_value = cls_loss
-                    .forward(class_preds, class_targets, Reduction::Mean)
-                    .mul_scalar(self.classification_weight)
-                    .mul_scalar(self.global_scale);
-
-                total_loss = total_loss + cls_loss_value;
-            }
+        if let Some(ref cls_loss) = self.classification_loss
+            && let (Some(class_preds), Some(class_targets)) = (class_preds, class_targets)
+        {
+            let cls_loss_value = cls_loss
+                .forward(class_preds, class_targets, Reduction::Mean)
+                .mul_scalar(self.classification_weight)
+                .mul_scalar(self.global_scale);
+
+            total_loss = total_loss + cls_loss_value;
         }
 
         Ok(total_loss)
@@ -314,21 +314,21 @@ impl<B: Backend> BiRefNetLoss<B> {
         let mut total_loss = pixel_loss;
 
         // Add classification loss if enabled and predictions are provided
-        if let Some(ref cls_loss) = self.classification_loss {
-            if let (Some(class_preds), Some(class_targets)) = (class_preds, class_targets) {
-                let cls_loss_value = cls_loss
-                    .forward(class_preds, class_targets, Reduction::Mean)
-                    .mul_scalar(self.classification_weight)
-                    .mul_scalar(self.global_scale);
-
-                total_loss = total_loss.clone() + cls_loss_value.clone();
-
-                // Add classification loss to dictionary
-                loss_dict.insert(
-                    "classification".to_owned(),
-                    cls_loss_value.into_scalar().to_f64(),
-                );
-            }
+        if let Some(ref cls_loss) = self.classification_loss
+            && let (Some(class_preds), Some(class_targets)) = (class_preds, class_targets)
+        {
+            let cls_loss_value = cls_loss
+                .forward(class_preds, class_targets, Reduction::Mean)
+                .mul_scalar(self.classification_weight)
+                .mul_scalar(self.global_scale);
+
+            total_loss = total_loss.clone() + cls_loss_value.clone();
+
+            // Add classification loss to dictionary
+            loss_dict.insert(
+                "classification".to_owned(),
+                cls_loss_value.into_scalar().to_f64(),
+            );
         }
 
         // Add total loss to dictionary

crates/birefnet-loss/src/lib.rs

@@ -106,4 +106,4 @@ mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}

crates/birefnet-metric/src/biou.rs

@@ -4,10 +4,9 @@
 //! for evaluating boundary-focused segmentation performance.
 
 use core::marker::PhantomData;
-
-use birefnet_util::{StructuringElement, erosion};
 use std::sync::Arc;
 
+use birefnet_util::{StructuringElement, erosion};
 use burn::{
     prelude::*,
     tensor::{Tensor, backend::Backend, cast::ToElement},

crates/birefnet-metric/src/lib.rs

@@ -96,15 +96,20 @@ pub use mse::{MSEMetric, MSEMetricConfig};
 // Re-export lesser-used items with warning
 #[deprecated(note = "S-measure is not yet fully implemented and tested")]
 #[doc(inline)]
-pub use s_measure::{SMeasureInput as SMeasureInputDeprecated, SMeasureMetric, SMeasureMetricConfig, calculate_s_measure};
+pub use s_measure::{
+    SMeasureInput as SMeasureInputDeprecated, SMeasureMetric, SMeasureMetricConfig,
+    calculate_s_measure,
+};
 #[doc(inline)]
 pub use utils::{AllMetricsResult, calculate_all_metrics};
 #[doc(inline)]
-pub use weighted_f_measure::{WeightedFMeasureMetric, WeightedFMeasureMetricConfig, calculate_weighted_f_measure};
+pub use weighted_f_measure::{
+    WeightedFMeasureMetric, WeightedFMeasureMetricConfig, calculate_weighted_f_measure,
+};
 
 #[cfg(test)]
 mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}

crates/birefnet-model/src/lib.rs

@@ -49,4 +49,4 @@ mod tests {
     pub type TestBackend = Cpu;
 
     pub type TestAutodiffBackend = Autodiff<TestBackend>;
-}
+}

crates/birefnet-train/src/dataset.rs

@@ -78,10 +78,9 @@ impl<B: Backend> Batcher<B, BiRefNetItem, BiRefNetBatch<B>> for BiRefNetBatcher<
 
             // Apply ImageNet normalization: add batch dimension, normalize, then remove batch dimension
             let image_tensor_with_batch = image_tensor.unsqueeze::<4>(); // [C, H, W] -> [1, C, H, W]
-            let normalized_tensor =
-                apply_imagenet_normalization(image_tensor_with_batch)
-                    .expect("Failed to apply ImageNet normalization")
-                    .squeeze::<3>(); // [1, C, H, W] -> [C, H, W]
+            let normalized_tensor = apply_imagenet_normalization(image_tensor_with_batch)
+                .expect("Failed to apply ImageNet normalization")
+                .squeeze::<3>(); // [1, C, H, W] -> [C, H, W]
 
             // Convert mask data to tensor [1, H, W]
             let mask_tensor = Tensor::<B, 2>::from_data(

crates/birefnet-train/src/lib.rs

@@ -15,4 +15,4 @@ mod tests {
     use burn::backend::Cpu;
 
     pub type TestBackend = Cpu;
-}
+}