!8507 [MSLITE] Check validation of input shapes before running graphes.

From: @wang_shaocong
Reviewed-by: @zhang_xue_tong
Signed-off-by:

mindspore/lite/nnacl/split_parameter.h

@@ -24,7 +24,7 @@ typedef struct SplitParameter {
   OpParameter op_parameter_;
   SplitQuantArg quant_arg_;
   int num_split_;
-  int split_sizes_[32];
+  int *split_sizes_;
   int strides_[32];
   int split_dim_;
   int n_dims_;

mindspore/lite/src/executor.cc

@@ -29,6 +29,13 @@ int Executor::CheckInputs(std::vector<Tensor *> &in_tensors) {
       MS_LOG(ERROR) << "Graph input tensor data is nullptr";
       return RET_ERROR;
     }
+    auto shape = inTensor->shape();
+    bool valid = all_of(shape.begin(), shape.end(), [](int i) { return i > 0; });
+    if (!valid) {
+      MS_LOG(ERROR) << "The shape of input tensor contains zero or negative dimension,"
+                    << "check the model and assign the input shape with method Resize().";
+      return RET_ERROR;
+    }
   }
   return RET_OK;
 }

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

@@ -32,13 +32,19 @@ OpParameter *PopulateSplitParameter(const mindspore::lite::PrimitiveC *primitive
   auto param = reinterpret_cast<mindspore::lite::Split *>(const_cast<mindspore::lite::PrimitiveC *>(primitive));
   split_param->op_parameter_.type_ = primitive->Type();
   split_param->num_split_ = param->GetNumberSplit();
+  int *split_sizes = reinterpret_cast<int *>(malloc(split_param->num_split_ * sizeof(int)));
+  if (split_sizes == nullptr) {
+    MS_LOG(ERROR) << "malloc split size of SplitParameter failed.";
+    return nullptr;
+  }
+  memset(split_sizes, 0, split_param->num_split_ * sizeof(int));
+  split_param->split_sizes_ = split_sizes;
   auto split_sizes_vector_ = param->GetSizeSplits();
   int i = 0;
   for (auto iter = split_sizes_vector_.begin(); iter != split_sizes_vector_.end(); iter++) {
     split_param->split_sizes_[i++] = *iter;
   }
   split_param->split_dim_ = param->GetSplitDim();
-  split_param->num_split_ = param->GetNumberSplit();
   return reinterpret_cast<OpParameter *>(split_param);
 }
 Registry SplitParameterRegistry(schema::PrimitiveType_Split, PopulateSplitParameter);

mindspore/lite/tools/converter/legacy_optimizer/graph/trans_format_insert_pass.cc

@@ -294,7 +294,10 @@ STATUS TransOpInsertPass::Run(schema::MetaGraphT *graph) {
         return ret;
       }
       ret = ChangeOpAxis(graph, node);
-      if (ret != RET_OK) {
+      if (ret == RET_NOT_SUPPORT) {
+        MS_LOG(INFO) << "not support to ChangeOpAxis";
+        return RET_OK;
+      } else if (ret != RET_OK) {
         MS_LOG(INFO) << "no need to ChangeOpAxis";
         return ret;
       }

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

@@ -251,6 +251,7 @@ STATUS OnnxModelParser::ParseOnnxGivenFillNode(const onnx::NodeProto &onnx_node)
             quant_param->zeroPoint = static_cast<int32_t>(onnx_node_attr.i());
           }
         }
+        quant_param->inited = true;
         tensor->quantParams.emplace_back(std::move(quant_param));
       } else {
         MS_LOG(ERROR) << "unsupported data type " << tensor->dataType;
@@ -369,6 +370,7 @@ void OnnxModelParser::SetOpQuantParams(const onnx::GraphProto &onnx_graph, const
       MS_LOG(ERROR) << "new QuantParamT failed, node: " << dst_op->name;
       return;
     }
+    quant_param->inited = true;
     int argNum = 0;
     for (const auto &onnx_node_attr : node.attribute()) {
       if (onnx_node_attr.name() == "Y_scale") {
@@ -384,6 +386,7 @@ void OnnxModelParser::SetOpQuantParams(const onnx::GraphProto &onnx_graph, const
       quant_param->zeroPoint = 0;
       quant_param->min = FLT_MAX;
       quant_param->max = FLT_MAX;
+      quant_param->inited = false;
     }
     dst_tensor->quantParams.emplace_back(std::move(quant_param));
     if (argNum == 2) {

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

@@ -39,9 +39,9 @@ STATUS OnnxQuantizeParser::Parse(const onnx::GraphProto &onnx_graph, const onnx:
   }
   if (onnx_node.op_type() == "Int8Quantize") {
     attr->srcT = kNumberTypeFloat32;
-    attr->dstT = kNumberTypeInt8;
+    attr->dstT = kNumberTypeUInt8;
   } else if (onnx_node.op_type() == "Int8Dequantize") {
-    attr->srcT = kNumberTypeInt8;
+    attr->srcT = kNumberTypeUInt8;
     attr->dstT = kNumberTypeFloat32;
   } else {
     MS_LOG(ERROR) << "Unsupported nodeType: " << onnx_node.op_type().c_str();