!12392 [MSLITE] Fix the bug of fp16 gather working in multithreading

From: @zhanyuan1
Reviewed-by: @zhanghaibo5,@zhang_xue_tong
Signed-off-by: @zhang_xue_tong

mindspore/lite/src/runtime/kernel/arm/fp16/gather_fp16.cc

@@ -91,9 +91,6 @@ int GatherFp16CPUKernel::DoGather(int task_id) {
   auto thread_stride = stride * task_id;
   int8_t *int8_in = nullptr;
   if (input_tensor->data_type() == kNumberTypeFloat32) {
-    input_data_ =
-      reinterpret_cast<float16_t *>(context_->allocator->Malloc(input_tensor->ElementsNum() * sizeof(float16_t)));
-    Float32ToFloat16(reinterpret_cast<float *>(input_tensor->data_c()), input_data_, input_tensor->ElementsNum());
     int8_in = reinterpret_cast<int8_t *>(input_data_);
   } else if (input_tensor->data_type() == kNumberTypeFloat16) {
     int8_in = reinterpret_cast<int8_t *>(input_tensor->data_c());
@@ -127,7 +124,12 @@ int GatherFp16CPUKernel::Run() {
     MS_LOG(ERROR) << "AssignIndicesData failed, error_code[" << ret << "]";
     return ret;
   }
-
+  auto input_tensor = in_tensors_.at(0);
+  if (input_tensor->data_type() == kNumberTypeFloat32) {
+    input_data_ =
+      reinterpret_cast<float16_t *>(context_->allocator->Malloc(input_tensor->ElementsNum() * sizeof(float16_t)));
+    Float32ToFloat16(reinterpret_cast<float *>(input_tensor->data_c()), input_data_, input_tensor->ElementsNum());
+  }
   ret = ParallelLaunch(this->context_->thread_pool_, GatherRunFp16, this, op_parameter_->thread_num_);
   if (ret != RET_OK) {
     MS_LOG(ERROR) << "Gather function error error_code[" << ret << "]";
@@ -140,7 +142,6 @@ int GatherFp16CPUKernel::Run() {
     context_->allocator->Free(input_data_);
     input_data_ = nullptr;
   }
-
   return ret;
 }