[TensorRT EP] Add null_ptr check to avoid crash when running session which was failed to generate trt_engine previously (#21621)

### Description
<!-- Describe your changes. -->
Add null_ptr check to avoid crash when running session which was failed
to generate trt_engine previously


### Motivation and Context
<!-- - Why is this change required? What problem does it solve?
- If it fixes an open issue, please link to the issue here. -->

Reported and verified by
https://github.com/microsoft/onnxruntime/issues/21567

onnxruntime/core/providers/tensorrt/tensorrt_execution_provider.cc

@@ -3752,6 +3752,11 @@ Status TensorrtExecutionProvider::CreateNodeComputeInfoFromGraph(const GraphView
       trt_context = trt_state->context->get();
     }
 
+    // Check before using trt_engine
+    if (trt_engine == nullptr) {
+      return ORT_MAKE_STATUS(ONNXRUNTIME, EP_FAIL, "No engine is found.");
+    }
+
     // Get input and output binding names
     int total_bindings = trt_engine->getNbIOTensors();
     std::vector<char const*> input_binding_names, output_binding_names;
@@ -4075,6 +4080,11 @@ Status TensorrtExecutionProvider::CreateNodeComputeInfoFromPrecompiledEngine(con
     Ort::ThrowOnError(api->KernelContext_GetGPUComputeStream(context, &cuda_stream));
     cudaStream_t stream = static_cast<cudaStream_t>(cuda_stream);
 
+    // Check before using trt_engine
+    if (trt_engine == nullptr) {
+      return ORT_MAKE_STATUS(ONNXRUNTIME, EP_FAIL, "No engine is found.");
+    }
+
     // Get input and output binding names
     int total_bindings = trt_engine->getNbIOTensors();
     std::vector<char const*> input_binding_names, output_binding_names;