fix reduce relu6

mindspore/lite/nnacl/fp16/activation_fp16.c

@@ -34,28 +34,20 @@ int ReluFp16(const float16_t *src, float16_t *dst, int ele_num) {
 }
 
 int Relu6Fp16(const float16_t *data, float16_t *dst, int ele_num) {
-  int eight_block = UP_DIV(ele_num, C8NUM);
-  int i;
-  for (i = 0; i < eight_block - 1; i++) {
-    int index = i * C8NUM;
+  int offset = 0;
 #ifdef ENABLE_NEON
-    float16x8_t relu6_data = vld1q_f16(data + index);
-    float16x8_t zero_data = vdupq_n_f16(0);
-    float16x8_t six_data = vdupq_n_f16(6);
+  float16x8_t zero_data = vdupq_n_f16(0);
+  float16x8_t six_data = vdupq_n_f16(6);
+  for (; offset <= ele_num - C8NUM; offset += C8NUM) {
+    float16x8_t relu6_data = vld1q_f16(data + offset);
     relu6_data = vmaxq_f16(relu6_data, zero_data);
     relu6_data = vminq_f16(relu6_data, six_data);
-    vst1q_f16(dst + index, relu6_data);
-#else
-    int j;
-    for (j = 0; j < C8NUM; ++j) {
-      dst[index + j] = data[index + j] < 0 ? 0 : data[index + j];
-      dst[index + j] = dst[index + j] > 6 ? 6 : dst[index + j];
-    }
-#endif
+    vst1q_f16(dst + offset, relu6_data);
   }
-  for (int j = (eight_block - 1) * C8NUM; j < ele_num; ++j) {
-    dst[j] = data[j] < 0 ? 0 : data[j];
-    dst[j] = dst[j] > 6 ? 6 : dst[j];
+#endif
+  for (; offset < ele_num; offset++) {
+    dst[offset] = data[offset] < 0 ? 0 : data[offset];
+    dst[offset] = dst[offset] > 6 ? 6 : dst[offset];
   }
   return NNACL_OK;
 }

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

@@ -82,14 +82,7 @@ int ReduceCPUKernel::Init() {
   return ReSize();
 }
 
-int ReduceCPUKernel::ReSize() {
-  if (in_tensors().at(0)->data_type() == kNumberTypeFloat32) {
-    data_type_ = kDataTypeFloat;
-  } else {
-    data_type_ = kDataTypeInt;
-  }
-  return ReduceBaseCPUKernel::ReSize();
-}
+int ReduceCPUKernel::ReSize() { return ReduceBaseCPUKernel::ReSize(); }
 
 int ReduceCPUKernel::CallReduceUnit(int task_id) {
   int ret;
@@ -120,6 +113,11 @@ int ReduceCPUKernel::Run() {
     MS_LOG(ERROR) << "Prepare fail!ret: " << prepare_ret;
     return prepare_ret;
   }
+  if (in_tensors().at(0)->data_type() == kNumberTypeFloat32) {
+    data_type_ = kDataTypeFloat;
+  } else {
+    data_type_ = kDataTypeInt;
+  }
   auto ret = MallocTmpBuffer();
   if (ret != RET_OK) {
     FreeTmpBuffer();