Optimize ORT graph execution and Linux webcam latency

- Enable ORT_ENABLE_BASIC graph optimization instead of fully disabling it, yielding CPU fallback performance improvements without crashing the GPU on FusedConv.

- Set cv2.CAP_PROP_BUFFERSIZE to 1 on Linux VideoCapture to eliminate webcam tracking lag/latency.

input_reader.py

@@ -15,6 +15,7 @@ class VideoReader():
             self.cap = cv2.VideoCapture(capture, cv2.CAP_DSHOW)
         else:
             self.cap = cv2.VideoCapture(capture)
+            self.cap.set(cv2.CAP_PROP_BUFFERSIZE, 1)
         if self.cap is None:
             print("The video source cannot be opened")
             sys.exit(0)

retinaface.py

@@ -64,7 +64,7 @@ class RetinaFaceDetector():
         options.inter_op_num_threads = 1
         options.intra_op_num_threads = threads
         options.execution_mode = onnxruntime.ExecutionMode.ORT_SEQUENTIAL
-        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_DISABLE_ALL
+        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_ENABLE_BASIC
         options.log_severity_level = 3
         providersList = ['CPUExecutionProvider']
         self.session = onnxruntime.InferenceSession(model_path, sess_options=options, providers=providersList)

tracker.py

@@ -499,7 +499,7 @@ class Tracker():
         options.inter_op_num_threads = 1
         options.intra_op_num_threads = min(max_threads,4)
         options.execution_mode = onnxruntime.ExecutionMode.ORT_SEQUENTIAL
-        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_DISABLE_ALL
+        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_ENABLE_BASIC
         options.log_severity_level = 3
         self.model_type = model_type
         self.models = [
@@ -547,7 +547,7 @@ class Tracker():
             elif i < extra_threads:
                 options.intra_op_num_threads += 1
             options.execution_mode = onnxruntime.ExecutionMode.ORT_SEQUENTIAL
-            options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_DISABLE_ALL
+            options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_ENABLE_BASIC
             self.sessions.append(onnxruntime.InferenceSession(os.path.join(model_base_path, model), sess_options=options, providers=providersList))
         self.input_name = self.session.get_inputs()[0].name
 
@@ -555,7 +555,7 @@ class Tracker():
         options.inter_op_num_threads = 1
         options.intra_op_num_threads = 1
         options.execution_mode = onnxruntime.ExecutionMode.ORT_SEQUENTIAL
-        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_DISABLE_ALL
+        options.graph_optimization_level = onnxruntime.GraphOptimizationLevel.ORT_ENABLE_BASIC
         options.log_severity_level = 3
         self.gaze_model = onnxruntime.InferenceSession(os.path.join(model_base_path, "mnv3_gaze32_split_gpu.onnx"), sess_options=options, providers=providersList)