diff --git a/go/demo/mobilenet_c.cc b/go/demo/mobilenet_c.cc
index e20635b0bb7bc617731fa8181ca15555192bb1be..6a5cc683c9f9a9c88f73a3ca5ebac274210f3b7a 100644
--- a/go/demo/mobilenet_c.cc
+++ b/go/demo/mobilenet_c.cc
@@ -29,6 +29,7 @@ int main(int argc, char *argv[]) {
   printf("Output num: %d\n", output_num);
 
   PD_ZeroCopyTensor input;
+  PD_InitZeroCopyTensor(&input);
   input.name = const_cast<char *>(PD_GetInputName(predictor, 0));  // NOLINT
   input.data.capacity = sizeof(float) * 1 * 3 * 300 * 300;
   input.data.length = input.data.capacity;
@@ -48,6 +49,7 @@ int main(int argc, char *argv[]) {
   PD_InitZeroCopyTensor(&output);
   output.name = const_cast<char *>(PD_GetOutputName(predictor, 0));  // NOLINT
   PD_GetZeroCopyOutput(predictor, &output);
+
   PD_DestroyZeroCopyTensor(&output);
 
   PD_DeleteAnalysisConfig(config);
diff --git a/go/demo/mobilenet_cxx.cc b/go/demo/mobilenet_cxx.cc
index a3df9a0e930bfe8755899f39f7d868dd1438f55f..7bdd6b2b03b24e2393e746edde754f763e9dd986 100644
--- a/go/demo/mobilenet_cxx.cc
+++ b/go/demo/mobilenet_cxx.cc
@@ -35,13 +35,13 @@ int main(int argc, char *argv[]) {
   input->copy_from_cpu(data.data());
   predictor->ZeroCopyRun();
   auto output_name = predictor->GetOutputNames()[0];
-  output = predictor->GetOutputTensor(output_name);
+  auto output = predictor->GetOutputTensor(output_name);
   return 0;
 }
 
 void SetConfig(paddle::AnalysisConfig *config) {
   config->SetModel("data/model/__model__", "data/model/__params__");
-  config->SwitchUseFeedFetchOps(true);
+  config->SwitchUseFeedFetchOps(false);
   config->SwitchSpecifyInputNames(true);
   config->SwitchIrOptim(false);
 }
diff --git a/go/paddle/common.go b/go/paddle/common.go
index b29efbdf3022b644d42b22cc5f005f12e0775ab9..346dc3f395bfe6ba4ae9fdc716f41a58b2bb4725 100644
--- a/go/paddle/common.go
+++ b/go/paddle/common.go
@@ -21,6 +21,14 @@ package paddle
 import "C"
 import "fmt"
 
+type Precision C.Precision
+
+const (
+	kFloat32 Precision = C.kFloat32
+	kInt8    Precision = C.kInt8
+	kHalf    Precision = C.kHalf
+)
+
 func ConvertCBooleanToGo(b C.bool) bool {
 	var c_false C.bool
 	if b != c_false {
diff --git a/go/paddle/config.go b/go/paddle/config.go
index 524249a067121f0094ca6c215e96aa85963c8132..5480365d58069b44b24fbb027e62e7809e4eb5cf 100644
--- a/go/paddle/config.go
+++ b/go/paddle/config.go
@@ -43,8 +43,13 @@ func (config *AnalysisConfig) SetModel(model, params string) {
 	//C.printString((*C.char)(unsafe.Pointer(&s[0])))
 	c_model := C.CString(model)
 	defer C.free(unsafe.Pointer(c_model))
-	c_params := C.CString(params)
-	defer C.free(unsafe.Pointer(c_params))
+	var c_params *C.char
+	if params == "" {
+		c_params = nil
+	} else {
+		c_params = C.CString(params)
+		defer C.free(unsafe.Pointer(c_params))
+	}
 
 	C.PD_SetModel(config.c, c_model, c_params)
 }
@@ -61,8 +66,8 @@ func (config *AnalysisConfig) ParamsFile() string {
 	return C.GoString(C.PD_ParamsFile(config.c))
 }
 
-func (config *AnalysisConfig) EnableUseGpu(memory_pool_init_size_mb uint64, device_id int) {
-	C.PD_EnableUseGpu(config.c, C.ulong(memory_pool_init_size_mb), C.int(device_id))
+func (config *AnalysisConfig) EnableUseGpu(memory_pool_init_size_mb int, device_id int) {
+	C.PD_EnableUseGpu(config.c, C.int(memory_pool_init_size_mb), C.int(device_id))
 }
 
 func (config *AnalysisConfig) DisableGpu() {
@@ -113,7 +118,9 @@ func (config *AnalysisConfig) SpecifyInputName() bool {
 	return ConvertCBooleanToGo(C.PD_SpecifyInputName(config.c))
 }
 
-//func (config *AnalysisConfig) EnableTensorRtEngine(workspace_size int)
+func (config *AnalysisConfig) EnableTensorRtEngine(workspace_size int, max_batch_size int, min_subgraph_size int, precision Precision, use_static bool, use_calib_mode bool) {
+	C.PD_EnableTensorRtEngine(config.c, C.int(workspace_size), C.int(max_batch_size), C.int(min_subgraph_size), C.Precision(precision), C.bool(use_static), C.bool(use_calib_mode))
+}
 
 func (config *AnalysisConfig) TensorrtEngineEnabled() bool {
 	return ConvertCBooleanToGo(C.PD_TensorrtEngineEnabled(config.c))
@@ -175,15 +182,15 @@ func (config *AnalysisConfig) DisableGlogInfo() {
 }
 
 func (config *AnalysisConfig) DeletePass(pass string) {
-    c_pass := C.CString(pass)
-    defer C.free(unsafe.Pointer(c_pass))
-    C.PD_DeletePass(config.c, c_pass)
+	c_pass := C.CString(pass)
+	defer C.free(unsafe.Pointer(c_pass))
+	C.PD_DeletePass(config.c, c_pass)
 }
 
 func (config *AnalysisConfig) SetInValid() {
-    C.PD_SetInValid(config.c)
+	C.PD_SetInValid(config.c)
 }
 
 func (config *AnalysisConfig) IsValid() bool {
-    return ConvertCBooleanToGo(C.PD_IsValid(config.c))
+	return ConvertCBooleanToGo(C.PD_IsValid(config.c))
 }
diff --git a/paddle/fluid/inference/capi/paddle_c_api.h b/paddle/fluid/inference/capi/paddle_c_api.h
index 87fb515dd6a6a25125c170188cce792951f0ccfd..e9157c06dec3659e3287d8157b6b29a10a7830da 100644
--- a/paddle/fluid/inference/capi/paddle_c_api.h
+++ b/paddle/fluid/inference/capi/paddle_c_api.h
@@ -161,9 +161,9 @@ PADDLE_CAPI_EXPORT extern const char* PD_ProgFile(
 PADDLE_CAPI_EXPORT extern const char* PD_ParamsFile(
     const PD_AnalysisConfig* config);
 
-PADDLE_CAPI_EXPORT extern void PD_EnableUseGpu(
-    PD_AnalysisConfig* config, uint64_t memory_pool_init_size_mb,
-    int device_id);
+PADDLE_CAPI_EXPORT extern void PD_EnableUseGpu(PD_AnalysisConfig* config,
+                                               int memory_pool_init_size_mb,
+                                               int device_id);
 
 PADDLE_CAPI_EXPORT extern void PD_DisableGpu(PD_AnalysisConfig* config);
 
diff --git a/paddle/fluid/inference/capi/pd_config.cc b/paddle/fluid/inference/capi/pd_config.cc
index 368489fde29cc0f51798e8ec7aebf413628c3403..458637377364f92fe976a90fbe7d3e953aed53da 100644
--- a/paddle/fluid/inference/capi/pd_config.cc
+++ b/paddle/fluid/inference/capi/pd_config.cc
@@ -79,10 +79,11 @@ const char* PD_ParamsFile(const PD_AnalysisConfig* config) {
   return config->config.params_file().c_str();
 }
 
-void PD_EnableUseGpu(PD_AnalysisConfig* config,
-                     uint64_t memory_pool_init_size_mb, int device_id) {
+void PD_EnableUseGpu(PD_AnalysisConfig* config, int memory_pool_init_size_mb,
+                     int device_id) {
   PADDLE_ENFORCE_NOT_NULL(config);
-  config->config.EnableUseGpu(memory_pool_init_size_mb, device_id);
+  config->config.EnableUseGpu(static_cast<uint64_t>(memory_pool_init_size_mb),
+                              device_id);
 }
 
 void PD_DisableGpu(PD_AnalysisConfig* config) {