Fixes

Author: nikita-kud

src/plugins/intel_npu/src/backend/src/zero_device.cpp

@@ -196,8 +196,8 @@ std::pair<std::unordered_map<std::string, std::shared_ptr<ov::ITensor>>, ov::SoP
                                                      nullptr,
                                                      inputTensorsData,
                                                      outputTensorsData,
-                                                     /*numberOfCommandLists*/ 1,
                                                      groupOrdinal);
+
     begin = std::chrono::steady_clock::now();
     pipeline->push();
     pipeline->pull();

src/plugins/intel_npu/src/backend/src/zero_pipeline.cpp

@@ -66,7 +66,7 @@ Pipeline::Pipeline(const Config& config,
             }
 
             graph->set_argument_value(desc.idx,
-                                      static_cast<unsigned char*>(inputTensorsData.at(ioIndex).at(0)->mem) +
+                                      static_cast<const unsigned char*>(inputTensorsData.at(ioIndex).at(0)->mem) +
                                           (i * inputTensorsData.at(ioIndex).at(0)->size) / _number_of_command_lists);
 
             ++ioIndex;
@@ -75,7 +75,7 @@ Pipeline::Pipeline(const Config& config,
         ioIndex = 0;
         for (const auto& desc : graph->get_output_descriptors()) {
             graph->set_argument_value(desc.idx,
-                                      static_cast<unsigned char*>(outputTensorsData.at(ioIndex)->mem) +
+                                      static_cast<const unsigned char*>(outputTensorsData.at(ioIndex)->mem) +
                                           (i * outputTensorsData.at(ioIndex)->size) / _number_of_command_lists);
             ++ioIndex;
         }
@@ -189,7 +189,7 @@ void Pipeline::updateCommandList(const TensorData& tensorsData, uint32_t index)
     for (size_t i = 0; i < _number_of_command_lists; i++) {
         _command_lists.at(i)->updateMutableCommandList(
             index,
-            static_cast<unsigned char*>(tensorsData.mem) + (i * tensorsData.size) / _number_of_command_lists);
+            static_cast<const unsigned char*>(tensorsData.mem) + (i * tensorsData.size) / _number_of_command_lists);
         _command_lists.at(i)->close();
     }
 };

src/plugins/intel_npu/src/common/include/intel_npu/common/igraph.hpp

@@ -97,8 +97,6 @@ class IGraph : public std::enable_shared_from_this<IGraph> {
     // first inference starts running
     std::mutex _mutex;
 
-    std::vector<uint8_t> _blob;
-
     uint32_t _unique_id = 0;
     uint32_t _last_submitted_id;