!11716 [MS][LITE]fix bug of tensorliset when elementshape is nullptr

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

mindspore/lite/src/ops/tensorlist_setitem.cc

@@ -131,8 +131,13 @@ int TensorListSetItem::InferShape(std::vector<lite::Tensor *> inputs_, std::vect
   }
 
   output0->set_max_elements_num(input0->max_elements_num());
-  output0->set_element_shape(input0->element_shape());
 
+  if (input0->tensors().empty() && input0->element_shape().empty() && index == 0) {
+    input0->set_element_shape(value_tensor->shape());
+    output0->set_element_shape(value_tensor->shape());
+  } else {
+    output0->set_element_shape(input0->element_shape());
+  }
   std::vector<std::vector<int> > out_shape;
   if (index == 0 && input0->tensors().size() == 0) {  // uninitialized tensorlist
     out_shape.push_back(value_tensor->shape());

mindspore/lite/src/runtime/kernel/arm/base/carry_data.cc

@@ -89,7 +89,9 @@ int CarryDataKernel::MoveTensorLiteData(lite::TensorList *dst_tensor, lite::Tens
     MS_LOG(ERROR) << "input tensorlist and output tensorlist data_type or format is incompatible";
     return RET_ERROR;
   }
-  if (dst_tensor->element_shape() != src_tensor->element_shape()) {
+  if (dst_tensor->element_shape().empty()) {
+    dst_tensor->set_element_shape(src_tensor->element_shape());
+  } else if (dst_tensor->element_shape() != src_tensor->element_shape()) {
     MS_LOG(ERROR) << "input tensorlist and output tensorlist element shape is incompatible";
     return RET_ERROR;
   }

mindspore/lite/src/runtime/kernel/arm/fp32/tensorlist_getitem_fp32.cc

@@ -68,7 +68,7 @@ int TensorListGetItemCPUKernel::Run() {
   } else {
     // reset 0 and dtype = dtype_
     // TODO(DT_VARIANT): dtype = DT_VARIANT is not handle
-    auto out_data = out_tensors_[0]->MutableData();
+    auto out_data = out_tensors_[0]->data_c();
     if (out_data == nullptr) {
       MS_LOG(ERROR) << "data of out_tensors_[0] is nullptr";
       return RET_ERROR;

mindspore/lite/src/runtime/kernel/arm/fp32/tensorlist_setitem_fp32.cc

@@ -30,6 +30,22 @@ namespace mindspore::kernel {
 
 int TensorListSetItemCPUKernel::Init() { return RET_OK; }
 
+int TensorListSetItemCPUKernel::CheckParam() {
+  if (dtype_ != kTypeUnknown && dtype_ != input0_->tensors_data_type()) {
+    MS_LOG(ERROR) << "op dtype:" << dtype_ << " is not equal in_tensors[0] dtype:" << input0_->data_type();
+    return RET_ERROR;
+  }
+  if (in_tensors_[1]->data_type() != kNumberTypeInt && in_tensors_[1]->data_type() != kNumberTypeInt32) {
+    MS_LOG(ERROR) << "in_tensors_[1]->data_type():" << in_tensors_[1]->data_type() << " must be int";
+    return RET_ERROR;
+  }
+  if (in_tensors_[1]->ElementsNum() != 1) {
+    MS_LOG(ERROR) << "in_tensors_[1]->ElementsNum():" << in_tensors_[1]->ElementsNum() << " must be equal to 1!";
+    return RET_ERROR;
+  }
+  return RET_OK;
+}
+
 int TensorListSetItemCPUKernel::IncrementOutputSize(int origin_size) {
   output0_ = reinterpret_cast<lite::TensorList *>(out_tensors_[0]);
   int new_tensors_size = origin_size + 1;
@@ -46,19 +62,13 @@ int TensorListSetItemCPUKernel::IncrementOutputSize(int origin_size) {
 
 int TensorListSetItemCPUKernel::Run() {
   input0_ = reinterpret_cast<lite::TensorList *>(in_tensors_[0]);
-  if (dtype_ != kTypeUnknown && dtype_ != input0_->tensors_data_type()) {
-    MS_LOG(ERROR) << "op dtype:" << dtype_ << " is not equal in_tensors[0] dtype:" << input0_->data_type();
+
+  if (CheckParam() != RET_OK) {
+    MS_LOG(ERROR) << "check param failed.";
     return RET_ERROR;
   }
+
   int dim0 = input0_->ElementsNum() - 1;
-  if (in_tensors_[1]->data_type() != kNumberTypeInt && in_tensors_[1]->data_type() != kNumberTypeInt32) {
-    MS_LOG(ERROR) << "in_tensors_[1]->data_type():" << in_tensors_[1]->data_type() << " must be int";
-    return RET_ERROR;
-  }
-  if (in_tensors_[1]->ElementsNum() != 1) {
-    MS_LOG(ERROR) << "in_tensors_[1]->ElementsNum():" << in_tensors_[1]->ElementsNum() << " must be equal to 1!";
-    return RET_ERROR;
-  }
   index_ = reinterpret_cast<int *>(in_tensors_[1]->data_c())[0];
   if (index_ < 0 || index_ > dim0) {
     if (IncrementOutputSize(output0_->shape()[0]) != RET_OK) {
@@ -81,6 +91,10 @@ int TensorListSetItemCPUKernel::Run() {
     }
   }
   // copy each tensor in tensors_
+  if (input0_->tensors().empty() && index_ == 0) {
+    input0_->set_element_shape(input2_->shape());
+    output0_->set_element_shape(input2_->shape());
+  }
   for (int i = 0; i < output0_->ElementsNum(); ++i) {
     if (i == index_) {
       auto dst = output0_->GetTensor(i);

mindspore/lite/src/runtime/kernel/arm/fp32/tensorlist_setitem_fp32.h

@@ -39,6 +39,7 @@ class TensorListSetItemCPUKernel : public LiteKernel {
   int IncrementOutputSize(int origin_size);
 
  private:
+  int CheckParam();
   lite::TensorList *input0_ = nullptr;
   lite::Tensor *input2_ = nullptr;
   lite::TensorList *output0_ = nullptr;

mindspore/lite/src/tensorlist.cc

@@ -240,6 +240,9 @@ Tensor *TensorList::GetTensor(int index) {
 }
 
 bool TensorList::IsCompatibleShape(const std::vector<int> &shape) {
+  if (this->tensors_.empty() && this->element_shape_.empty()) {
+    return true;
+  }
   if (shape.size() != this->element_shape_.size()) {
     return false;
   }

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

@@ -439,6 +439,9 @@ STATUS OnnxModelParser::BuildCNode(const onnx::NodeProto &onnx_node, const FuncG
               MS_LOG(ERROR) << "memcpy error: " << ret;
               return RET_ERROR;
             }
+            copy_param_value->set_tensor_shape(param_value->tensor_shape());
+            copy_param_value->set_format(param_value->format());
+            copy_param_value->set_tensor_type(param_value->tensor_type());
             copy_param_value->SetTensorData(copy_data, param_value->tensor_size());
             ext_subgraph_input->set_default_param(copy_param_value);
           } else {