!13087 fix bug for dynamic_batch_size

From: @zhupuxu
Reviewed-by: @zhoufeng54,@jjfeing
Signed-off-by: @jjfeing

mindspore/ccsrc/cxx_api/graph/acl/model_process.cc

@@ -53,15 +53,18 @@ inline static void PushbackIfNotNull(U *vec, T &&item) {
 }
 
 static void ConstructTensorDesc(const std::vector<AclTensorInfo> &acl_tensor_list, std::vector<std::string> *names,
-                                std::vector<std::vector<int64_t>> *shapes, std::vector<enum DataType> *data_types) {
+                                std::vector<std::vector<int64_t>> *shapes, std::vector<enum DataType> *data_types,
+                                std::vector<size_t> *mem_sizes) {
   ClearIfNotNull(names);
   ClearIfNotNull(shapes);
   ClearIfNotNull(data_types);
+  ClearIfNotNull(mem_sizes);
   for (size_t i = 0; i < acl_tensor_list.size(); ++i) {
     const auto &info = acl_tensor_list[i];
     PushbackIfNotNull(names, info.name);
     PushbackIfNotNull(shapes, info.dims);
     PushbackIfNotNull(data_types, TransToApiType(info.data_type));
+    PushbackIfNotNull(mem_sizes, info.buffer_size);
   }
 }
 
@@ -78,17 +81,20 @@ static std::string ShapeToString(const std::vector<int64_t> &shape) {
 }
 
 Status ModelProcess::ConstructTensors(const std::vector<AclTensorInfo> &acl_tensor_list,
-                                      std::vector<MSTensor> *tensor_list, const std::vector<size_t> &mem_sizes) {
+                                      std::vector<MSTensor> *tensor_list) {
   MS_EXCEPTION_IF_NULL(tensor_list);
   std::vector<std::string> names;
   std::vector<std::vector<int64_t>> shapes;
   std::vector<enum DataType> data_types;
-  ConstructTensorDesc(acl_tensor_list, &names, &shapes, &data_types);
+  std::vector<size_t> mem_sizes;
+
+  ConstructTensorDesc(acl_tensor_list, &names, &shapes, &data_types, &mem_sizes);
   tensor_list->clear();
   if (names.size() != acl_tensor_list.size() || shapes.size() != acl_tensor_list.size() ||
-      data_types.size() != acl_tensor_list.size()) {
+      data_types.size() != acl_tensor_list.size() || mem_sizes.size() != acl_tensor_list.size()) {
     MS_LOG(ERROR) << "Inner error, size do not match: names size " << names.size() << " shapes size " << shapes.size()
-                  << " data types size " << data_types.size() << " acl_tensor_list size " << acl_tensor_list.size();
+                  << " data types size " << data_types.size() << " mem sizes size " << mem_sizes.size()
+                  << " acl_tensor_list size " << acl_tensor_list.size();
     return kMCFailed;
   }
 
@@ -96,7 +102,7 @@ Status ModelProcess::ConstructTensors(const std::vector<AclTensorInfo> &acl_tens
   for (size_t i = 0; i < acl_tensor_list.size(); ++i) {
     tensor_list->emplace_back(names[i], data_types[i], shapes[i], nullptr, mem_sizes[i]);
     auto ret = aclrtMemcpy((*tensor_list)[i].MutableData(), (*tensor_list)[i].DataSize(),
-                           acl_tensor_list[i].device_data, mem_sizes[i], kind);
+                           acl_tensor_list[i].device_data, acl_tensor_list[i].buffer_size, kind);
     if (ret != ACL_ERROR_NONE) {
       MS_LOG(ERROR) << "Memcpy input " << i << " from " << (is_run_on_device_ ? "host" : "device")
                     << " to host failed, memory size " << acl_tensor_list[i].buffer_size;
@@ -159,7 +165,6 @@ Status ModelProcess::InitInputsBuffer() {
     }
     MS_LOG(INFO) << "Name of input " << i << " is " << input_name;
     input_infos_.emplace_back(AclTensorInfo{data_mem_buffer, buffer_size, data_type, shape, input_name});
-    input_size_.push_back(buffer_size);
   }
   MS_LOG(INFO) << "Create model inputs success";
   return kSuccess;
@@ -328,9 +333,8 @@ size_t ModelProcess::GetDynamicDims(const std::vector<AclTensorInfo> &inputs) {
 Status ModelProcess::SetBatchSize(const std::vector<MSTensor> &inputs) {
   size_t index;
   aclError ret;
-  input_size_.clear();
-  for (auto input : inputs) {
-    input_size_.push_back(input.DataSize());
+  for (size_t i = 0; i < inputs.size(); i++) {
+    input_infos_[i].buffer_size = inputs[i].DataSize();
   }
   auto *p = reinterpret_cast<const float *>(inputs[inputs.size() - 1].Data().get());
   MS_EXCEPTION_IF_NULL(p);
@@ -353,12 +357,12 @@ Status ModelProcess::CheckAndInitInput(const std::vector<MSTensor> &inputs) {
   inputs_ = aclmdlCreateDataset();
   size_t dynamic_nums = GetDynamicDims(input_infos_);
   // check inputs
+  if (inputs.size() != input_infos_.size()) {
+    MS_LOG(ERROR) << "Inputs count not match, required count " << input_infos_.size() << ", given count "
+                  << inputs.size();
+    return kMCInvalidInput;
+  }
   if (dynamic_nums == 0) {
-    if (inputs.size() != input_infos_.size()) {
-      MS_LOG(ERROR) << "Inputs count not match, required count " << input_infos_.size() << ", given count "
-                    << inputs.size();
-      return kMCInvalidInput;
-    }
     for (size_t i = 0; i < input_infos_.size(); ++i) {
       if (inputs[i].Shape() != input_infos_[i].dims) {
         MS_LOG(INFO) << "Note: input " << i << " shape not match, required " << ShapeToString(input_infos_[i].dims)
@@ -404,18 +408,18 @@ Status ModelProcess::CheckAndInitInput(const std::vector<MSTensor> &inputs) {
     }
   }
   if (dynamic_nums == 1) {
-    if (SetBatchSize(inputs) == kMCDeviceError) {
+    if (SetBatchSize(inputs) != kSuccess) {
       MS_LOG(ERROR) << "failed to convert dynamic batch size";
       return kMCDeviceError;
     }
+    if (ResetOutputSize() != kSuccess) {
+      MS_LOG(ERROR) << "reset output size failed";
+      return kMCDeviceError;
+    }
   } else if (dynamic_nums == 2) {
     MS_LOG(ERROR) << "only dynamic batch size is supported";
     return kMCInvalidInput;
   }
-  if (ResetOutputSize() == kMCDeviceError) {
-    MS_LOG(ERROR) << "reset output size failed";
-    return kMCDeviceError;
-  }
   return kSuccess;
 }
 
@@ -435,7 +439,7 @@ Status ModelProcess::ResetOutputSize() {
       dims *= output_dims.dims[i];
     }
     output_type = aclmdlGetOutputDataType(model_desc_, index);
-    output_size_.push_back(dims * aclDataTypeSize(output_type));
+    output_infos_[index].buffer_size = dims * aclDataTypeSize(output_type);
   }
   return kSuccess;
 }
@@ -490,7 +494,7 @@ Status ModelProcess::BuildOutputs(std::vector<MSTensor> *outputs) {
 }
 
 std::vector<MSTensor> ModelProcess::GetInputs() {
-  Status ret = ConstructTensors(input_infos_, &input_tensors_, input_size_);
+  Status ret = ConstructTensors(input_infos_, &input_tensors_);
   if (ret != kSuccess) {
     MS_LOG(ERROR) << "ConstructTensors failed.";
     input_tensors_.clear();
@@ -500,7 +504,7 @@ std::vector<MSTensor> ModelProcess::GetInputs() {
 }
 
 std::vector<MSTensor> ModelProcess::GetOutputs() {
-  Status ret = ConstructTensors(output_infos_, &output_tensors_, output_size_);
+  Status ret = ConstructTensors(output_infos_, &output_tensors_);
   if (ret != kSuccess) {
     MS_LOG(ERROR) << "ConstructTensors failed.";
     output_tensors_.clear();

mindspore/ccsrc/cxx_api/graph/acl/model_process.h

@@ -61,8 +61,7 @@ class ModelProcess {
  private:
   Status CreateDataBuffer(void **data_mem_buffer, size_t buffer_size, aclmdlDataset *dataset);
   Status CheckAndInitInput(const std::vector<MSTensor> &inputs);
-  Status ConstructTensors(const std::vector<AclTensorInfo> &acl_tensor_list, std::vector<MSTensor> *tensor_list,
-                          const std::vector<size_t> &mem_sizes);
+  Status ConstructTensors(const std::vector<AclTensorInfo> &acl_tensor_list, std::vector<MSTensor> *tensor_list);
   Status BuildOutputs(std::vector<MSTensor> *outputs);
   Status SetBatchSize(const std::vector<MSTensor> &inputs);
   Status InitInputsBuffer();
@@ -84,8 +83,6 @@ class ModelProcess {
   std::vector<AclTensorInfo> output_infos_;
   std::vector<MSTensor> input_tensors_;
   std::vector<MSTensor> output_tensors_;
-  std::vector<size_t> output_size_;
-  std::vector<size_t> input_size_;
   size_t GetDynamicDims(const std::vector<AclTensorInfo> &);
 };
 }  // namespace mindspore