Fixed linting

docker/perception/semantic_segmentation/semantic_segmentation.Dockerfile

@@ -18,14 +18,14 @@ WORKDIR ${AMENT_WS}/src
 RUN git clone -b main https://github.com/open-mmlab/mmsegmentation.git /mmsegmentation
 WORKDIR /mmsegmentation
 ENV FORCE_CUDA="1"
-#RUN pip install -r requirements.txt
 RUN pip install mmengine
 RUN pip install --no-cache-dir -e .
 RUN ["/bin/bash", "-c", "pip install openmim"]
 
 # Scan for dependencies
 RUN echo "openmim\nmmengine\nmmcv==2.0.1\n" > /tmp/pip_install_list.txt
 
+# Install segformer model
 RUN mim download mmsegmentation --config segformer_mit-b2_8xb1-160k_cityscapes-1024x1024 --dest ./model
 
 # # ################################ Source ################################
@@ -50,8 +50,6 @@ RUN apt-get -qq update && rosdep update && \
 ################################# Dependencies ################################
 FROM ${BASE_BUILD_IMAGE} as dependencies
 
-# # install tensorrt
-# RUN apt update && apt install -y tensorrt
 RUN apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/3bf863cc.pub
 RUN apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/7fa2af80.pub
 
@@ -68,7 +66,7 @@ RUN apt-get update && apt-get install -y \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
-
+# Install Segformer dependencies 
 COPY --from=Segformer /tmp/pip_install_list.txt /tmp/pip_install_list.txt
 RUN pip3 install torch==1.13.1+cu116 torchvision==0.14.1+cu116 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu116
 RUN pip install cython
@@ -109,7 +107,6 @@ COPY --from=Segformer /mmsegmentation /mmsegmentation
 RUN pip install -r requirements.txt
 RUN pip install --no-cache-dir -e .
 WORKDIR ${AMENT_WS}
-#  --from=Segformer /mmsegmentation/model /model
 # Add runtime libraries to path
 ENV CUDNN_DIR=/mmsegmentation/cuda
 ENV CV2_CUDABACKEND=0

src/perception/semantic_segmentation/launch/semantic_segmentation.launch.py

@@ -16,7 +16,7 @@ def generate_launch_description():
         executable='semantic_segmentation_node',
         name='semantic_segmentation_node',
         parameters=[config]
-        )
+    )
 
     # finalize
     ld.add_action(semantic_segmentation_node)

src/perception/semantic_segmentation/semantic_segmentation/segmentation_node.py

@@ -24,10 +24,9 @@
 package_name = 'semantic_segmentation'
 package_share_directory = get_package_share_directory(package_name)
 CONFIG = os.path.join(package_share_directory, 'resource', 'model',
-                       'segformer_mit-b2_8xb1-160k_cityscapes-1024x1024.py')
-CHECKPOINT = os.path.join(package_share_directory, 'resource', 'model', 
-                          'segformer_mit-b2_8x1_1024x1024_160k_cityscapes_20211207_134205-6096669a.pth')
-
+                    'segformer_mit-b2_8xb1-160k_cityscapes-1024x1024.py')
+CHECKPOINT = os.path.join(package_share_directory, 'resource', 'model',
+                'segformer_mit-b2_8x1_1024x1024_160k_cityscapes_20211207_134205-6096669a.pth')
 IMAGE_H = 900
 IMAGE_W = 1600
 
@@ -106,13 +105,13 @@ def listener_callback(self, msg):
                 except CvBridgeError as e:
                     self.get_logger().error(str(e))
                     return
-
         with torch.no_grad():
-            out_img = self.model(image, show=False, )['predictions']
-
-        # logits = torch.tensor(out_img, dtype=torch.float32).unsqueeze(0).unsqueeze(0)  # Add batch and channel dimensions
+            out_img = self.model(image, show=False, )['predictions']   
+        # logits = torch.tensor(
+        # out_img, dtype=torch.float32).unsqueeze(0).unsqueeze(0)
+        # Add batch and channel dimensions
         # upsampled_logits = torch.nn.functional.interpolate(logits,
-        #                                                    size=(IMAGE_H, IMAGE_W),  # (height, width)
+        #                    size=(IMAGE_H, IMAGE_W),  # (height, width)
         #                                                    mode='bilinear',
         #                                                    align_corners=False)
         # upsampled_logits = upsampled_logits.squeeze().numpy().astype(np.uint8)