fix gathernd bug

mindspore/lite/src/runtime/kernel/arm/fp32/gatherNd_fp32.cc

@@ -17,6 +17,7 @@
 #include "src/runtime/kernel/arm/fp32/gatherNd_fp32.h"
 #include <string.h>
 #include <limits>
+#include <vector>
 #include "schema/model_generated.h"
 #include "include/errorcode.h"
 #include "src/kernel_registry.h"
@@ -65,11 +66,17 @@ int GatherNdCPUKernel::ReSize() {
     return RET_ERROR;
   }
   (void)memset(in_offset_, 0, count_ * sizeof(int));
-
   thread_sz_count_ = MSMIN(thread_count_, count_);
   thread_sz_stride_ = UP_DIV(count_, thread_sz_count_);
+  return RET_OK;
+}
 
+void GatherNdCPUKernel::InitOffset() {
+  MS_ASSERT(in_offset_ != nullptr);
+  auto indices_tensor = in_tensors_.at(1);
+  auto indices_shape = indices_tensor->shape();
   auto in_shape = in_tensors_.front()->shape();
+  int indices_rank = indices_shape.size();
   int in_rank = in_shape.size();
   int idx_lastshape = indices_shape[indices_rank - 1];
   auto indices_ptr = reinterpret_cast<int *>(indices_tensor->MutableData());
@@ -89,8 +96,6 @@ int GatherNdCPUKernel::ReSize() {
       in_offset_[j] += indices_ptr[j * idx_stride + k] * in_stride.at(k);
     }
   }
-
-  return RET_OK;
 }
 
 int GatherNdCPUKernel::DoGatherNd(int task_id) {
@@ -120,6 +125,7 @@ int GatherNdRun(void *cdata, int task_id) {
 int GatherNdCPUKernel::Run() {
   in_ptr_ = reinterpret_cast<float *>(in_tensors_.front()->MutableData());
   out_ptr_ = reinterpret_cast<float *>(out_tensors_.front()->MutableData());
+  InitOffset();
   auto ret = ParallelLaunch(this->context_->thread_pool_, GatherNdRun, this, thread_sz_count_);
   if (ret != RET_OK) {
     MS_LOG(ERROR) << "gatherNd error error_code[" << ret << "]";

mindspore/lite/src/runtime/kernel/arm/fp32/gatherNd_fp32.h

@@ -41,6 +41,7 @@ class GatherNdCPUKernel : public LiteKernel {
   int DoGatherNd(int task_id);
 
  private:
+  void InitOffset();
   int thread_sz_count_;
   int thread_sz_stride_;
   int count_;

mindspore/lite/src/runtime/kernel/arm/int8/gatherNd_int8.cc

@@ -17,6 +17,7 @@
 #include "src/runtime/kernel/arm/int8/gatherNd_int8.h"
 #include <string.h>
 #include <limits>
+#include <vector>
 #include "schema/model_generated.h"
 #include "include/errorcode.h"
 #include "src/kernel_registry.h"
@@ -50,7 +51,6 @@ int GatherNdInt8CPUKernel::ReSize() {
     in_offset_ = nullptr;
   }
   auto in_quant_args = in_tensors_.at(0)->quant_params();
-  auto ind_quant_args = in_tensors_.at(1)->quant_params();
   auto out_quant_args = out_tensors_.at(0)->quant_params();
   param_.alpha_ = in_quant_args.front().scale / out_quant_args.front().scale;
   param_.zp_in_ = in_quant_args.front().zeroPoint;
@@ -73,10 +73,16 @@ int GatherNdInt8CPUKernel::ReSize() {
     return RET_ERROR;
   }
   (void)memset(in_offset_, 0, count_ * sizeof(int));
-
   thread_sz_count_ = MSMIN(thread_count_, count_);
   thread_sz_stride_ = UP_DIV(count_, thread_sz_count_);
+  return RET_OK;
+}
 
+void GatherNdInt8CPUKernel::InitOffset() {
+  auto ind_quant_args = in_tensors_.at(1)->quant_params();
+  auto indices_tensor = in_tensors_.at(1);
+  auto indices_shape = indices_tensor->shape();
+  int indices_rank = indices_shape.size();
   auto in_shape = in_tensors_.front()->shape();
   int in_rank = in_shape.size();
   int idx_lastshape = indices_shape.at(indices_rank - 1);
@@ -99,7 +105,6 @@ int GatherNdInt8CPUKernel::ReSize() {
       in_offset_[j] += tmp * in_stride[k];
     }
   }
-  return RET_OK;
 }
 
 int GatherNdInt8CPUKernel::DoGatherNd(int task_id) {
@@ -129,6 +134,7 @@ int GatherNdInt8Run(void *cdata, int task_id) {
 int GatherNdInt8CPUKernel::Run() {
   in_ptr_ = reinterpret_cast<int8_t *>(in_tensors_.front()->MutableData());
   out_ptr_ = reinterpret_cast<int8_t *>(out_tensors_.front()->MutableData());
+  InitOffset();
   auto ret = ParallelLaunch(this->context_->thread_pool_, GatherNdInt8Run, this, thread_sz_count_);
   if (ret != RET_OK) {
     MS_LOG(ERROR) << "gatherNd error error_code[" << ret << "]";

mindspore/lite/src/runtime/kernel/arm/int8/gatherNd_int8.h

@@ -36,6 +36,7 @@ class GatherNdInt8CPUKernel : public LiteKernel {
   int DoGatherNd(int task_id);
 
  private:
+  void InitOffset();
   int thread_count_;
   int thread_sz_count_;
   int thread_sz_stride_;