Add TensorRT execution provider support

This commit adds support for the ONNX Runtime TensorRT execution provider, including detection in CMake, new localization strings, and integration in background and enhance filters. It also updates session creation logic to handle TensorRT-specific options and cache management.

Author: umireon

CMakeLists.txt

@@ -183,6 +183,17 @@ if(HAVE_ONNXRUNTIME_ROCM_EP)
   target_compile_definitions(${CMAKE_PROJECT_NAME} PRIVATE HAVE_ONNXRUNTIME_ROCM_EP)
 endif()
 
+check_library_exists(
+  onnxruntime::onnxruntime
+  OrtSessionOptionsAppendExecutionProvider_Tensorrt
+  onnxruntime_c_api.h
+  HAVE_ONNXRUNTIME_TENSORRT_EP
+)
+if(HAVE_ONNXRUNTIME_TENSORRT_EP)
+  message(STATUS "ONNX Runtime TensorRT Execution Provider found")
+  target_compile_definitions(${CMAKE_PROJECT_NAME} PRIVATE HAVE_ONNXRUNTIME_TENSORRT_EP)
+endif()
+
 target_sources(
   ${CMAKE_PROJECT_NAME}
   PRIVATE

data/locale/ar-EG.ini

@@ -9,6 +9,7 @@ InferenceDevice="جهاز الإستدلال"
 CPU="وحدة المعالجة المركزية"
 GPUCUDA="الوحدة المركزية - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="نموذج التقسيم"
 SINet="SINet"

data/locale/bn-IN.ini

@@ -9,6 +9,7 @@ InferenceDevice="নিখরচনা ডিভাইস"
 CPU="সিপিইউ"
 GPUCUDA="জিপিইউ - কুড়া"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="কোরএমএল"
 SegmentationModel="সেগমেন্টেশন মডেল"
 SINet="এসআইনেট"

data/locale/en-US.ini

@@ -9,6 +9,7 @@ InferenceDevice="Inference device"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="Segmentation model"
 SINet="SINet"

data/locale/es-SP.ini

@@ -9,6 +9,7 @@ InferenceDevice="Dispositivo de inferencia"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="Modelo de segmentación"
 SINet="SINet"

data/locale/fr-FR.ini

@@ -9,6 +9,7 @@ InferenceDevice="Dispositif d'inférence"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="Modèle de segmentation"
 SINet="SINet"

data/locale/hi-IN.ini

@@ -9,6 +9,7 @@ InferenceDevice="संदर्भ डिवाइस"
 CPU="सीपीयू"
 GPUCUDA="जीपीयू - क्यूडा"
 GPUROCM="जीपीयू - ROCM"
+TENSORRT="TensorRT"
 CoreML="कोरएमएल"
 SegmentationModel="सेगमेंटेशन मॉडल"
 SINet="ऐसआईनेट"

data/locale/it-IT.ini

@@ -9,6 +9,7 @@ InferenceDevice="Dispositivo di inferenza"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="Modello di segmentazione".
 SINet="SINet"

data/locale/ja-JP.ini

@@ -9,6 +9,7 @@ InferenceDevice="推論デバイス"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="セグメンテーションモデル"
 SINet="SINet"

data/locale/ko-KR.ini

@@ -9,6 +9,7 @@ InferenceDevice="추론 장치"
 CPU="CPU"
 GPUCUDA="GPU - CUDA"
 GPUROCM="GPU - ROCM"
+TENSORRT="TensorRT"
 CoreML="CoreML"
 SegmentationModel="세분화 모델"
 SINet="SINet"