Add config for TensorRT and CUDA execution provider (#992)

Signed-off-by: manickavela1998@gmail.com <manickavela1998@gmail.com>
Signed-off-by: manickavela1998@gmail.com <manickavela.arumugam@uniphore.com>

sherpa-onnx/python/csrc/cuda-config.cc

@@ -0,0 +1,24 @@
+// sherpa-onnx/python/csrc/cuda-config.cc
+//
+// Copyright (c)  2024  Uniphore (Author: Manickavela A)
+
+#include "sherpa-onnx/python/csrc/cuda-config.h"
+
+#include <memory>
+#include <string>
+
+#include "sherpa-onnx/csrc/provider-config.h"
+
+namespace sherpa_onnx {
+
+void PybindCudaConfig(py::module *m) {
+  using PyClass = CudaConfig;
+  py::class_<PyClass>(*m, "CudaConfig")
+      .def(py::init<>())
+      .def(py::init<int32_t>(),
+           py::arg("cudnn_conv_algo_search") = 1)
+      .def_readwrite("cudnn_conv_algo_search", &PyClass::cudnn_conv_algo_search)
+      .def("__str__", &PyClass::ToString);
+}
+
+}  // namespace sherpa_onnx