!15277 fix fuzz bug & codex

From: @hangangqiang
Reviewed-by: @jpc_chenjianping,@zhanghaibo5
Signed-off-by: @zhanghaibo5

mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/fp32_grad/layernorm_grad.c

@@ -16,12 +16,16 @@
 #include "nnacl/fp32_grad/layernorm_grad.h"
 #include <stddef.h>
 #include <math.h>
+#include "nnacl/errorcode.h"
 
-void LayerNormGrad(const float *x, const float *dy, const float *var, const float *mean, const float *gamma,
-                   int param_num, int param_size, int block_num, int block_size, float *dx, float *dg, float *db) {
+int LayerNormGrad(const float *x, const float *dy, const float *var, const float *mean, const float *gamma,
+                  int param_num, int param_size, int block_num, int block_size, float *dx, float *dg, float *db) {
   // var is actually layer_norm forward output var
   const float eps = 1e-12;
   const float *var_sqrt_rev = var;
+  if (block_size <= 0) {
+    return NNACL_ERRCODE_DIVISOR_ZERO;
+  }
   for (size_t i = 0; i < param_num; ++i) {
     float dgamma = 0.0f;
     float dbeta = 0.0f;
@@ -56,4 +60,5 @@ void LayerNormGrad(const float *x, const float *dy, const float *var, const floa
       dx[index] = dx1 + dx2 + dx3;
     }
   }
+  return NNACL_OK;
 }

mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/fp32_grad/layernorm_grad.h

@@ -20,8 +20,8 @@
 extern "C" {
 #endif
 
-void LayerNormGrad(const float *x, const float *dy, const float *var, const float *mean, const float *gamma,
-                   int param_num, int param_size, int block_num, int block_size, float *dx, float *dg, float *db);
+int LayerNormGrad(const float *x, const float *dy, const float *var, const float *mean, const float *gamma,
+                  int param_num, int param_size, int block_num, int block_size, float *dx, float *dg, float *db);
 
 #ifdef __cplusplus
 }

mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/infer/arithmetic_infer.c

@@ -62,7 +62,6 @@ int ArithmeticInferShape(const TensorC *const *inputs, size_t inputs_size, Tenso
       }
       in_shape1_[i] = input_shape1[i];
     }
-    // format = input0->format();
   } else if (input_shape0_size > input_shape1_size) {
     ndim_ = input_shape0_size;
     int fill_dim_num = input_shape0_size - input_shape1_size;

mindspore/ccsrc/backend/kernel_compiler/cpu/nnacl/infer/strided_slice_grad_infer.c

@@ -89,7 +89,7 @@ int StridedSliceGradInferShape(const TensorC *const *inputs, size_t inputs_size,
     param->strides_[i] = strides_[i];
   }
   ShapeSet(param->in_shape_, &in_shape_size, input->shape_, input->shape_size_);
-  // ApplyNewAxisMask();
+  // ApplyNewAxisMask;
   for (size_t i = 0; i < ndim_; i++) {
     if (new_axis_mask_[i]) {
       ndim_ += 1;
@@ -107,19 +107,19 @@ int StridedSliceGradInferShape(const TensorC *const *inputs, size_t inputs_size,
       ellipsis_mask_[i] = false;
     }
   }
-  // ApplyBeginMask();
+  // ApplyBeginMask;
   for (size_t i = 0; i < ndim_; i++) {
     if (begins_mask_[i]) {
       begins_[i] = 0;
     }
   }
-  // ApplyEndMask();
+  // ApplyEndMask;
   for (size_t i = 0; i < ndim_; i++) {
     if (ends_mask_[i]) {
       ends_[i] = in_shape_[i];
     }
   }
-  // ApplyEllipsisMask();
+  // ApplyEllipsisMask;
   for (size_t i = 0; i < ndim_; i++) {
     if (ellipsis_mask_[i]) {
       begins_[i] = 0;

mindspore/lite/examples/export_models/models/effnet.py

@@ -49,7 +49,6 @@ class Swish(nn.Cell):
         s = self.sigmoid(x)
         m = x*s
         return m
-        #return x * (1/(1+self.exp(-x)))
 
 
 class AdaptiveAvgPool(nn.Cell):
@@ -75,7 +74,6 @@ class SELayer(nn.Cell):
         self.act2 = nn.Sigmoid()
 
     def construct(self, x):
-        #b, c, _, _ = x.shape()
         o = self.avg_pool(x) #.view(b,c)
         o = self.conv_reduce(o)
         o = self.act1(o)
@@ -119,9 +117,6 @@ class DepthwiseSeparableConv(nn.Cell):
         x = self.bn2(x)
 
         if self.has_residual:
-            # if self.drop_connect_rate > 0.:
-                # x = x
-                # x = drop_connect(x, self.training, self.drop_connect_rate)
             x += residual
         return x
 
@@ -199,8 +194,6 @@ class InvertedResidual(nn.Cell):
         x = self.bn3(x)
 
         if self.has_residual:
-            # if self.drop_connect_rate > 0.:
-            #    x = x
             x += residual
         return x
 

mindspore/lite/examples/export_models/models/mobilenetv2_train_export.py

@@ -25,7 +25,6 @@ from mindspore.train.serialization import export
 context.set_context(mode=context.PYNATIVE_MODE, device_target="GPU", save_graphs=False)
 batch = 16
 
-#n = MobileNetV2()
 backbone_net = MobileNetV2Backbone()
 head_net = MobileNetV2Head(input_channel=backbone_net.out_channels, num_classes=10)
 n = mobilenet_v2(backbone_net, head_net)

mindspore/lite/examples/quick_start_cpp/build.sh

@@ -14,7 +14,7 @@
 # limitations under the License.
 # ============================================================================
 
-BASEPATH=$(cd "$(dirname $0)"; pwd)
+BASEPATH=$(cd "$(dirname $0)" || exit; pwd)
 get_version() {
     VERSION_MAJOR=$(grep "const int ms_version_major =" ${BASEPATH}/../../include/version.h | tr -dc "[0-9]")
     VERSION_MINOR=$(grep "const int ms_version_minor =" ${BASEPATH}/../../include/version.h | tr -dc "[0-9]")
@@ -39,6 +39,6 @@ fi
 tar xzvf ${BASEPATH}/build/${MINDSPORE_FILE} -C ${BASEPATH}/build/
 cp -r ${BASEPATH}/build/${MINDSPORE_FILE_NAME}/inference/lib/libmindspore-lite.a ${BASEPATH}/lib
 cp -r ${BASEPATH}/build/${MINDSPORE_FILE_NAME}/inference/include ${BASEPATH}/
-cd ${BASEPATH}/build
+cd ${BASEPATH}/build || exit
 cmake ${BASEPATH}
 make

mindspore/lite/examples/quick_start_java/build.sh

@@ -14,7 +14,7 @@
 # limitations under the License.
 # ============================================================================
 
-BASEPATH=$(cd "$(dirname $0)"; pwd)
+BASEPATH=$(cd "$(dirname $0)" || exit; pwd)
 get_version() {
     VERSION_MAJOR=$(grep "const int ms_version_major =" ${BASEPATH}/../../include/version.h | tr -dc "[0-9]")
     VERSION_MINOR=$(grep "const int ms_version_minor =" ${BASEPATH}/../../include/version.h | tr -dc "[0-9]")
@@ -38,6 +38,6 @@ if [ ! -e ${BASEPATH}/build/${MINDSPORE_FILE} ]; then
 fi
 tar xzvf ${BASEPATH}/build/${MINDSPORE_FILE} -C ${BASEPATH}/build/
 cp -r ${BASEPATH}/build/${MINDSPORE_FILE_NAME}/inference/lib/jar/* ${BASEPATH}/lib
-cd ${BASEPATH}/
+cd ${BASEPATH}/ || exit
 
 mvn package

mindspore/lite/examples/runtime_cpp/build.sh

@@ -24,7 +24,7 @@ usage()
 }
 
 BASEPATH=$(
-  cd "$(dirname $0)"
+  cd "$(dirname $0)" || exit
   pwd
 )
 get_version() {
@@ -94,7 +94,7 @@ cp -r ${BASEPATH}/build/${MINDSPORE_FILE_NAME}/inference/include ${BASEPATH}/
 if [[ "X${DEVICE}" == "Xnpu" ]]; then
     cp -r ${BASEPATH}/build/${MINDSPORE_FILE_NAME}/inference/third_party/hiai_ddk/lib/*.so ${BASEPATH}/lib
 fi
-cd ${BASEPATH}/build
+cd ${BASEPATH}/build || exit
 cmake -DCMAKE_TOOLCHAIN_FILE="${ANDROID_NDK}/build/cmake/android.toolchain.cmake" -DANDROID_NATIVE_API_LEVEL="19" \
   -DANDROID_NDK="${ANDROID_NDK}" -DANDROID_ABI="arm64-v8a" -DANDROID_STL="c++_shared" ${BASEPATH} -DSUPPORT_NPU=${SUPPORT_NPU}
 

mindspore/lite/src/common/tensor_util.cc

@@ -208,11 +208,11 @@ int GenerateInTensorC(const OpParameter *const parameter, const std::vector<lite
 int CheckTensorsInvalid(const std::vector<Tensor *> &tensors) {
   for (auto &tensor : tensors) {
     if (tensor == nullptr) {
-      MS_LOG(ERROR) << "check tensor is nullptr";
+      MS_LOG(ERROR) << "Graph input tensor is nullptr";
       return RET_ERROR;
     }
     if (tensor->data_type() != kObjectTypeTensorType && tensor->data_c() == nullptr) {
-      MS_LOG(ERROR) << "check tensor data is nullptr " << tensors;
+      MS_LOG(ERROR) << "Graph input tensor is nullptr " << tensors;
       return RET_ERROR;
     }
     auto shape = tensor->shape();

mindspore/lite/src/lite_model.cc

@@ -255,7 +255,29 @@ int LiteModel::SubGraphVerify() const {
   return RET_OK;
 }
 
-bool LiteModel::ModelVerify() const { return NodeVerify() == RET_OK && SubGraphVerify() == RET_OK; }
+bool LiteModel::ModelVerify() const {
+  if (this->sub_graphs_.empty()) {
+    MS_LOG(ERROR) << "Model does not have a main graph.";
+    return false;
+  }
+  auto main_graph = this->sub_graphs_.front();
+  for (auto input_index : main_graph->input_indices_) {
+    if (input_index >= this->all_tensors_.size()) {
+      MS_LOG(ERROR) << "Graph indices is beyond tensor_size.";
+      return false;
+    }
+    auto *tensor = this->all_tensors_.at(input_index);
+    if (tensor == nullptr) {
+      MS_LOG(ERROR) << "Tensor in all tensors is nullptr.";
+      return false;
+    }
+    if (tensor->format() != schema::Format_NHWC) {
+      MS_LOG(ERROR) << "Graph input tensor should be NHWC";
+      return false;
+    }
+  }
+  return NodeVerify() == RET_OK && SubGraphVerify() == RET_OK;
+}
 
 const void *LiteModel::GetMetaGraphByVerison() {
   MS_ASSERT(this->buf != nullptr);

mindspore/lite/src/lite_model.h

@@ -70,7 +70,11 @@ class LiteModel : public Model {
       } else if (node->quant_type_ == schema::QuantType_WeightQuant) {
         node->quant_type_ = schema::QuantType_QUANT_WEIGHT;
       }
-      node->name_ = c_node->name()->c_str();
+      if (c_node->name() == nullptr) {
+        node->name_ = "";
+      } else {
+        node->name_ = c_node->name()->c_str();
+      }
       auto count = c_node->inputIndex()->size();
       for (uint32_t j = 0; j < count; ++j) {
         node->input_indices_.push_back(size_t(c_node->inputIndex()->template GetAs<uint32_t>(j)));

mindspore/lite/src/ops/populate/depthwise_conv2d_populate.cc

@@ -26,47 +26,6 @@ OpParameter *PopulateConvDwParameter(const void *primitive) {
     return nullptr;
   }
   memset(conv_param, 0, sizeof(ConvParameter));
-  // conv_param->op_parameter_.type_ = primitive->Type();
-
-  // auto conv_primitive =
-  //   reinterpret_cast<mindspore::lite::DepthwiseConv2D *>(const_cast<mindspore::lite::PrimitiveC *>(primitive));
-  // conv_param->kernel_h_ = conv_primitive->GetKernelH();
-  // conv_param->kernel_w_ = conv_primitive->GetKernelW();
-  // conv_param->stride_h_ = conv_primitive->GetStrideH();
-  // conv_param->stride_w_ = conv_primitive->GetStrideW();
-
-  // auto convdw_lite_primitive = (lite::DepthwiseConv2D *)primitive;
-  // conv_param->pad_u_ = convdw_lite_primitive->PadUp();
-  // conv_param->pad_d_ = convdw_lite_primitive->PadDown();
-  // conv_param->pad_l_ = convdw_lite_primitive->PadLeft();
-  // conv_param->pad_r_ = convdw_lite_primitive->PadRight();
-  // conv_param->input_channel_ = convdw_lite_primitive->GetInputChannel();
-  // conv_param->dilation_h_ = conv_primitive->GetDilateH();
-  // conv_param->dilation_w_ = conv_primitive->GetDilateW();
-  // auto pad_mode = conv_primitive->GetPadMode();
-  // switch (pad_mode) {
-  //   case schema::PadMode_SAME_UPPER:
-  //     conv_param->pad_mode_ = Pad_Same;
-  //     break;
-  //   case schema::PadMode_VALID:
-  //     conv_param->pad_mode_ = Pad_Valid;
-  //     break;
-  //   default:
-  //     conv_param->pad_mode_ = Pad_No;
-  //     break;
-  // }
-  // auto act_type = conv_primitive->GetActivationType();
-  // switch (act_type) {
-  //   case schema::ActivationType_RELU:
-  //     conv_param->act_type_ = ActType_Relu;
-  //     break;
-  //   case schema::ActivationType_RELU6:
-  //     conv_param->act_type_ = ActType_Relu6;
-  //     break;
-  //   default:
-  //     conv_param->act_type_ = ActType_No;
-  //     break;
-  // }
   return reinterpret_cast<OpParameter *>(conv_param);
 }
 }  // namespace lite

mindspore/lite/src/ops/populate/mul_populate.cc

@@ -29,8 +29,6 @@ OpParameter *PopulateMulParameter(const void *prim) {
   }
   auto *primitive = static_cast<const schema::Primitive *>(prim);
   param->op_parameter_.type_ = primitive->value_type();
-  // auto mul_prim = primitive->value_as_Mul();
-  // param->activation_type_ = mul_prim->activationType();
   return reinterpret_cast<OpParameter *>(param);
 }
 }  // namespace

mindspore/lite/src/runtime/kernel/arm/fp32_grad/convolution.cc

@@ -115,8 +115,6 @@ int ConvolutionTrainCPUKernel::Execute(int task_id) {
         for (int j = 0; j < groups; ++j) {
           const float *mat_b = w_addr + j * nweights / groups;
           float *mat_c = y_addr + (i * groups) * n * m + j * (out_ch / groups) + ci * out_ch;
-          // float *im = x_addr + i * in_ch * in_h * in_w + j * (in_ch / groups);
-          // RollingIm2ColPackUnitFp32(im, conv_param_, mat_a, real_chunk, ci);
           GemmMatmul(0, 1, real_chunk, n, k, 1, mat_a + (j * kernel_spatial), k * groups, mat_b, k, 0, mat_c, out_ch,
                      mat_workspace);
         }

mindspore/lite/src/runtime/kernel/arm/fp32_grad/layernorm_grad.cc

@@ -21,6 +21,7 @@
 #include "src/kernel_registry.h"
 #include "nnacl/fp32_grad/layernorm_grad.h"
 #include "nnacl/fp32_grad/layernormgrad_parameter.h"
+#include "nnacl/errorcode.h"
 #include "include/errorcode.h"
 #include "src/runtime/runtime_api.h"
 
@@ -81,7 +82,11 @@ int LayerNormGradCPUKernel::Execute(int task_id) {
   float *dx = reinterpret_cast<float *>(output_dx->MutableData());
   float *dg = reinterpret_cast<float *>(output_dg->MutableData());
   float *db = reinterpret_cast<float *>(output_db->MutableData());
-  LayerNormGrad(x, dy, var, mean, gamma, param_num_, param_size_, block_num_, block_size_, dx, dg, db);
+  int ret = LayerNormGrad(x, dy, var, mean, gamma, param_num_, param_size_, block_num_, block_size_, dx, dg, db);
+  if (ret != NNACL_OK) {
+    MS_LOG(ERROR) << "LayerNormGrad error task_id[" << task_id << "] error_code[" << ret << "]";
+    return RET_ERROR;
+  }
   return RET_OK;
 }
 

mindspore/lite/src/runtime/kernel/opencl/kernel/conv2d.cc

@@ -278,7 +278,7 @@ void Conv2DOpenCLKernel::InitFilter() {
     allocator->UnmapBuffer(packed_filter_);
   }
 
-  FreeTmpWeight(in_tensors_.at(kWeightIndex)->data_c());
+  FreeTmpWeight(in_tensors_.at(kWeightIndex));
 }
 
 void Conv2DOpenCLKernel::InitBias() {
@@ -315,7 +315,7 @@ void Conv2DOpenCLKernel::InitBias() {
     }
   }
   allocator->UnmapBuffer(packed_bias_);
-  FreeTmpWeight(in_tensors_.at(kBiasIndex)->data_c());
+  FreeTmpWeight(in_tensors_.at(kBiasIndex));
 }
 
 void Conv2DOpenCLKernel::SetConstArgs() {

mindspore/lite/src/runtime/kernel/opencl/utils.cc

@@ -300,7 +300,7 @@ static std::set<void *> tmp_weights;
 
 void StoreTmpWeight(lite::Tensor *tensor) {
   MS_LOG(WARNING) << "store weight when kernel don't infer shape!";
-  if (tensor != nullptr && tensor->data_c() != nullptr && tensor->Size() > 0) {
+  if ((tensor != nullptr) && (tensor->data_c() != nullptr) && (tensor->Size() > 0)) {
     void *new_data = malloc(tensor->Size());
     MS_ASSERT(new_data);
     if (new_data == nullptr) {
@@ -312,10 +312,13 @@ void StoreTmpWeight(lite::Tensor *tensor) {
   }
 }
 
-void FreeTmpWeight(void *data) {
+void FreeTmpWeight(lite::Tensor *tensor) {
+  MS_ASSERT(tensor != nullptr);
+  auto data = tensor->data_c();
   if (tmp_weights.count(data)) {
     tmp_weights.erase(data);
     free(data);
+    tensor->set_data(nullptr);
   }
 }
 

mindspore/lite/src/runtime/kernel/opencl/utils.h

@@ -64,7 +64,7 @@ int CheckParamLikeTensor(const std::string &kernel_name, const std::string &tens
                          TypeId expect_data_type, const std::vector<int> &expect_shape);
 
 void StoreTmpWeight(lite::Tensor *tensor);
-void FreeTmpWeight(void *tensor);
+void FreeTmpWeight(lite::Tensor *tensor);
 
 template <class T1, class T2>
 void PackNCHWToNC4HW4(void *src, void *dst, int batch, int plane_in, int plane_out, int channel,

mindspore/lite/src/tensorlist.cc

@@ -142,6 +142,9 @@ int TensorList::MallocData(const mindspore::Allocator *allocator) {
 void TensorList::FreeData() {
   // free data buf of each tensor in tensors_
   for (auto tensor : tensors_) {
+    if (tensor == nullptr) {
+      continue;
+    }
     tensor->FreeData();
   }
 }
@@ -274,7 +277,11 @@ STATUS TensorList::Decode(const int *data) {
     element_shape_.push_back(data[2 + j]);
   }
   int tensors_num = data[2 + data[1]];
-  tensors_.resize(tensors_num);
+  if (this->ElementsNum() != tensors_num) {
+    MS_LOG(WARNING) << "Input tensorlist data is invalid: shape size(" << this->ElementsNum() << ") != tensors_num("
+                    << tensors_num << ").";
+  }
+  tensors_.reserve(tensors_num);
   int tensor_index = 2 + data[1] + 1;
   for (int i = 0; i < tensors_num; i++) {
     int tensor_dims_size = data[tensor_index++];
@@ -282,11 +289,12 @@ STATUS TensorList::Decode(const int *data) {
     for (int j = 0; j < tensor_dims_size; j++) {
       shape[j] = data[tensor_index++];
     }
-    tensors_[i] = new (std::nothrow) Tensor(tensors_data_type_, shape);
-    if (tensors_[i] == nullptr) {
+    auto tensor = new (std::nothrow) Tensor(tensors_data_type_, shape);
+    if (tensor == nullptr) {
       MS_LOG(ERROR) << "new Tensor failed";
       return RET_NULL_PTR;
     }
+    tensors_.emplace_back(tensor);
   }
   return RET_OK;
 }

mindspore/lite/tools/anf_exporter/anf_exporter.cc

@@ -753,7 +753,7 @@ int AnfExporter::ProcessTensor(const ValueNodePtr &value_node, std::unique_ptr<s
 
   // process weight tensor
   if (data->Size() > 0) {
-    if (memcpy_s((*schema_tensor)->data.data(), data->Size(), data->data_c(), data->Size()) != EOK) {
+    if (memcpy_s((*schema_tensor)->data.data(), (*schema_tensor)->data.size(), data->data_c(), data->Size()) != EOK) {
       MS_LOG(ERROR) << "memcpy_s error.";
       return RET_ERROR;
     }

mindspore/lite/tools/converter/converter_context.h

@@ -28,7 +28,6 @@ namespace mindspore {
 namespace lite {
 class ReturnCode {
  public:
-  virtual ~ReturnCode() = default;
   static ReturnCode *GetSingleReturnCode() {
     static ReturnCode return_code;
     return &return_code;
@@ -42,12 +41,12 @@ class ReturnCode {
 
  private:
   ReturnCode() = default;
+  virtual ~ReturnCode() = default;
   int status_code_ = RET_OK;
 };
 
 class NotSupportOp {
  public:
-  virtual ~NotSupportOp() = default;
   static NotSupportOp *GetInstance() {
     static NotSupportOp not_support_op;
     return &not_support_op;
@@ -67,13 +66,13 @@ class NotSupportOp {
 
  private:
   NotSupportOp() = default;
+  virtual ~NotSupportOp() = default;
   std::set<std::string> not_support_ops_;
   std::string fmk_type_;
 };
 
 class TensorDataType {
  public:
-  ~TensorDataType() = default;
   static TensorDataType *GetInstance() {
     static TensorDataType tensor_data_type;
     return &tensor_data_type;
@@ -88,6 +87,7 @@ class TensorDataType {
 
  private:
   TensorDataType() {}
+  virtual ~TensorDataType() = default;
   std::map<int32_t, int32_t> tensor_data_type_map_;
 };
 }  // namespace lite

mindspore/lite/tools/converter/parser/caffe/caffe_node_parser_registry.h

@@ -25,15 +25,16 @@
 namespace mindspore::lite {
 class CaffeNodeParserRegistry {
  public:
-  CaffeNodeParserRegistry();
-
-  virtual ~CaffeNodeParserRegistry();
-
   static CaffeNodeParserRegistry *GetInstance();
 
   CaffeNodeParser *GetNodeParser(const std::string &name);
 
   std::unordered_map<std::string, CaffeNodeParser *> parsers;
+
+ private:
+  CaffeNodeParserRegistry();
+
+  virtual ~CaffeNodeParserRegistry();
 };
 
 class CaffeNodeRegistrar {

mindspore/lite/tools/converter/parser/onnx/onnx_model_parser.cc

@@ -1127,7 +1127,7 @@ STATUS OnnxModelParser::CopyOnnxTensorData(const onnx::TensorProto &onnx_const_t
     return RET_MEMORY_FAILED;
   }
   auto tensor_data = reinterpret_cast<uint8_t *>(tensor_info->data_c());
-  if (memcpy_s(tensor_data, data_size, onnx_data, data_size) != EOK) {
+  if (memcpy_s(tensor_data, tensor_info->data().nbytes(), onnx_data, data_size) != EOK) {
     MS_LOG(ERROR) << "memcpy_s failed";
     return RET_ERROR;
   }

mindspore/lite/tools/converter/parser/onnx/onnx_node_parser_registry.h

@@ -25,15 +25,16 @@ namespace mindspore {
 namespace lite {
 class OnnxNodeParserRegistry {
  public:
-  OnnxNodeParserRegistry();
-
-  virtual ~OnnxNodeParserRegistry();
-
   static OnnxNodeParserRegistry *GetInstance();
 
   OnnxNodeParser *GetNodeParser(const std::string &name);
 
   std::unordered_map<std::string, OnnxNodeParser *> parsers;
+
+ private:
+  OnnxNodeParserRegistry();
+
+  virtual ~OnnxNodeParserRegistry();
 };
 
 class OnnxNodeRegistrar {

mindspore/lite/tools/converter/parser/tf/tf_node_parser_registry.h

@@ -25,14 +25,16 @@ namespace mindspore {
 namespace lite {
 class TFNodeParserRegistry {
  public:
-  TFNodeParserRegistry() = default;
-
-  virtual ~TFNodeParserRegistry();
-
   static TFNodeParserRegistry *GetInstance();
+
   TFNodeParser *GetNodeParser(const std::string &name);
 
   std::unordered_map<std::string, TFNodeParser *> parsers;
+
+ private:
+  TFNodeParserRegistry() = default;
+
+  virtual ~TFNodeParserRegistry();
 };
 
 class TFNodeRegistrar {

mindspore/lite/tools/converter/parser/tflite/tflite_node_parser_registry.h

@@ -25,15 +25,16 @@ namespace mindspore {
 namespace lite {
 class TfliteNodeParserRegistry {
  public:
-  TfliteNodeParserRegistry();
-
-  virtual ~TfliteNodeParserRegistry();
-
   static TfliteNodeParserRegistry *GetInstance();
 
   TfliteNodeParser *GetNodeParser(const tflite::BuiltinOperator &type);
 
   std::unordered_map<tflite::BuiltinOperator, TfliteNodeParser *> parsers;
+
+ private:
+  TfliteNodeParserRegistry();
+
+  virtual ~TfliteNodeParserRegistry();
 };
 
 class TfliteNodeRegister {

mindspore/lite/tools/optimizer/fusion/conv_transform_fusion.cc

@@ -52,6 +52,7 @@ int GetOutChannels(const CNodePtr &conv_node) {
 
 void GenerateNewWeightConv2D(float *dst_weight, const float *conv_weight, const float *scale_weight, FmkType fmk,
                              int weight_shape_size, int kernel_num) {
+  MS_ASSERT(kernel_num > 0);
   if (dst_weight == nullptr || conv_weight == nullptr || scale_weight == nullptr) {
     return;
   }
@@ -60,7 +61,6 @@ void GenerateNewWeightConv2D(float *dst_weight, const float *conv_weight, const
       dst_weight[i] = conv_weight[i] * scale_weight[i % kernel_num];
     }
   } else {
-    MS_ASSERT(kernel_num > 0);
     auto kernel_size = weight_shape_size / kernel_num;
     for (int i = 0; i < kernel_num; i++) {
       for (int j = 0; j < kernel_size; j++) {
@@ -75,6 +75,7 @@ void GenerateNewWeightConv2DTranspose(float *dst_weight, const float *scale_weig
   if (dst_weight == nullptr || scale_weight == nullptr || weight_tensor == nullptr) {
     return;
   }
+  MS_ASSERT(group > 0);
   auto weight_data = reinterpret_cast<float *>(weight_tensor->data_c());
   if (fmk == lite::converter::FmkType_TF) {
     auto cin_group = weight_tensor->shape()[3] / group;

model_zoo/official/lite/MindSpore_inhand/imageObject/src/main/cpp/SceneMindSporeNetnative.cpp

@@ -66,7 +66,6 @@ std::string SceneProcessRunnetResult(const int RET_CATEGORY_SUM,
     // Get a pointer to the first score.
     float *temp_scores = static_cast<float *>(outputTensor->MutableData());
 
-    // float scores[RET_CATEGORY_SUM];
     float scores = temp_scores[0];
     int cat_loc = 0;
     for (int i = 0; i < RET_CATEGORY_SUM; ++i) {

model_zoo/official/lite/MindSpore_inhand/segmentation/src/main/java/com/mindspore/imagesegmentation/StyleRecyclerViewAdapter.java

@@ -46,9 +46,6 @@ public class StyleRecyclerViewAdapter extends RecyclerView.Adapter<StyleRecycler
 
     @Override
     public void onBindViewHolder(@NonNull StyleItemViewHolder holder, int position) {
-//        Glide.with(context).
-//                load(IMAGES[position]).
-//                into(holder.getImageView());
         holder.imageView.setImageResource(IMAGES[position]);
         holder.itemView.setTag(IMAGES[position]);
         holder.itemView.setOnClickListener(view1 -> {

model_zoo/official/lite/MindSpore_inhand/segmentation/src/main/java/com/mindspore/imagesegmentation/help/TrackingMobile.java

@@ -111,7 +111,6 @@ public class TrackingMobile {
         ByteBuffer contentArray = BitmapUtils.bitmapToByteBuffer(scaledBitmap, imageSize, imageSize, IMAGE_MEAN, IMAGE_STD);
 
         MSTensor inTensor = inputs.get(0);
-        //    int byteLen = (int) inTensor.size();
         inTensor.setData(contentArray);
 
         // Run graph to infer results.

model_zoo/official/lite/image_segmentation/app/src/main/java/com/mindspore/imagesegmentation/help/TrackingMobile.java

@@ -111,7 +111,6 @@ public class TrackingMobile {
         ByteBuffer contentArray = BitmapUtils.bitmapToByteBuffer(scaledBitmap, imageSize, imageSize, IMAGE_MEAN, IMAGE_STD);
 
         MSTensor inTensor = inputs.get(0);
-        //    int byteLen = (int) inTensor.size();
         inTensor.setData(contentArray);
 
         // Run graph to infer results.

model_zoo/official/lite/scene_detection/app/src/main/cpp/MindSporeNetnative.cpp

@@ -235,7 +235,6 @@ std::string ProcessRunnetResult(const int RET_CATEGORY_SUM, const char *const la
     // Get a pointer to the first score.
     float *temp_scores = static_cast<float *>(outputTensor->MutableData());
 
-    // float scores[RET_CATEGORY_SUM];
     float scores = temp_scores[0];
     int cat_loc = 0;
     for (int i = 0; i < RET_CATEGORY_SUM; ++i) {

model_zoo/official/lite/style_transfer/app/src/main/java/com/mindspore/styletransferdemo/StyleTransferModelExecutor.java

@@ -127,20 +127,16 @@ public class StyleTransferModelExecutor {
     }
 
     @SuppressLint("LongLogTag")
-//    public ModelExecutionResult execute(String contentImagePath, String styleImageName) {
     public ModelExecutionResult execute(Bitmap contentImage, Bitmap styleBitmap) {
         Log.i(TAG, "running models");
 
         fullExecutionTime = SystemClock.uptimeMillis();
         preProcessTime = SystemClock.uptimeMillis();
 
-//        Bitmap contentImage = ImageUtils.decodeBitmap(new File(contentImagePath));
         ByteBuffer contentArray =
                 ImageUtils.bitmapToByteBuffer(contentImage, CONTENT_IMAGE_SIZE, CONTENT_IMAGE_SIZE, 0, 255);
 
 
-//        Bitmap styleBitmap =
-//                ImageUtils.loadBitmapFromResources(context, "thumbnails/" + styleImageName);
         ByteBuffer input = ImageUtils.bitmapToByteBuffer(styleBitmap, STYLE_IMAGE_SIZE, STYLE_IMAGE_SIZE, 0, 255);