!13582 【MD】【BUG】fix MD lite train error

From: @xulei2020
Reviewed-by: 
Signed-off-by:

mindspore/ccsrc/minddata/dataset/core/data_type.h

@@ -257,10 +257,12 @@ inline DataType DataType::FromCType<float>() {
   return DataType(DataType::DE_FLOAT32);
 }
 
+#ifndef ENABLE_MD_LITE_X86_64
 template <>
 inline DataType DataType::FromCType<float16>() {
   return DataType(DataType::DE_FLOAT16);
 }
+#endif
 
 template <>
 inline DataType DataType::FromCType<int64_t>() {
@@ -327,10 +329,12 @@ inline bool DataType::IsLooselyCompatible<float>() const {
   return type_ == DataType::DE_FLOAT32;
 }
 
+#ifndef ENABLE_MD_LITE_X86_64
 template <>
 inline bool DataType::IsLooselyCompatible<float16>() const {
   return type_ == DataType::DE_FLOAT16;
 }
+#endif
 
 template <>
 inline bool DataType::IsLooselyCompatible<int64_t>() const {

mindspore/ccsrc/minddata/dataset/core/tensor.cc

@@ -396,9 +396,9 @@ void Tensor::PrintItemAt(const std::vector<dsize_t> &index, std::ostream &out) c
     CASE_PRINT(DataType::DE_INT64, int64_t)
 
     CASE_PRINT(DataType::DE_UINT64, uint64_t)
-
+#ifndef ENABLE_MD_LITE_X86_64
     CASE_PRINT(DataType::DE_FLOAT16, float16)
-
+#endif
     CASE_PRINT(DataType::DE_FLOAT32, float)
 
     CASE_PRINT(DataType::DE_FLOAT64, double)
@@ -825,12 +825,14 @@ Status Tensor::GetFloatAt(T *o, const std::vector<dsize_t> &index) const {
     RETURN_STATUS_UNEXPECTED(err);
   }
   switch (type_.value()) {
+#ifndef ENABLE_MD_LITE_X86_64
     case DataType::DE_FLOAT16: {
       float16 *ptr = nullptr;
       RETURN_IF_NOT_OK(GetItemPtr<float16>(&ptr, index));
       *o = static_cast<T>(*ptr);
       break;
     }
+#endif
     case DataType::DE_FLOAT32: {
       float *ptr = nullptr;
       RETURN_IF_NOT_OK(GetItemPtr<float>(&ptr, index));

mindspore/ccsrc/minddata/dataset/kernels/data/data_utils.cc

@@ -281,9 +281,11 @@ void CastFrom(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *out
     case DataType::DE_UINT64:
       Cast<T, uint64_t>(input, output);
       break;
+#ifndef ENABLE_MD_LITE_X86_64
     case DataType::DE_FLOAT16:
       Cast<T, float16>(input, output);
       break;
+#endif
     case DataType::DE_FLOAT32:
       Cast<T, float>(input, output);
       break;
@@ -328,9 +330,11 @@ Status TypeCast(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *o
     case DataType::DE_UINT64:
       CastFrom<uint64_t>(input, output);
       break;
+#ifndef ENABLE_MD_LITE_X86_64
     case DataType::DE_FLOAT16:
       CastFrom<float16>(input, output);
       break;
+#endif
     case DataType::DE_FLOAT32:
       CastFrom<float>(input, output);
       break;
@@ -344,6 +348,7 @@ Status TypeCast(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *o
   return Status::OK();
 }
 
+#ifndef ENABLE_MD_LITE_X86_64
 Status ToFloat16(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *output) {
   // initiate new tensor for type cast
   DataType new_type = DataType("float16");
@@ -367,6 +372,9 @@ Status ToFloat16(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *
 
   return Status::OK();
 }
+#else
+Status ToFloat16(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *output) { return Status::OK(); }
+#endif
 
 Status PadEnd(const std::shared_ptr<Tensor> &src, std::shared_ptr<Tensor> *dst, const std::vector<dsize_t> &pad_shape,
               const std::shared_ptr<Tensor> &pad_val) {
@@ -410,9 +418,13 @@ Status PadEndNumeric(const std::shared_ptr<Tensor> &src, std::shared_ptr<Tensor>
       RETURN_IF_NOT_OK((*dst)->Fill<uint8_t>(pad_val));
     } else if (tensor_type == DataType::DE_INT16) {
       RETURN_IF_NOT_OK((*dst)->Fill<int16_t>(pad_val));
-    } else if (tensor_type == DataType::DE_FLOAT16) {
+    }
+#ifndef ENABLE_MD_LITE_X86_64
+    else if (tensor_type == DataType::DE_FLOAT16) {  // NOLINT
       RETURN_IF_NOT_OK((*dst)->Fill<float16>(static_cast<float16>(pad_val)));
-    } else if (tensor_type == DataType::DE_UINT16) {
+    }
+#endif
+    else if (tensor_type == DataType::DE_UINT16) {  // NOLINT
       RETURN_IF_NOT_OK((*dst)->Fill<uint16_t>(pad_val));
     } else if (tensor_type == DataType::DE_INT32) {
       RETURN_IF_NOT_OK((*dst)->Fill<int32_t>(pad_val));
@@ -570,9 +582,11 @@ Status Mask(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *outpu
     case DataType::DE_INT64:
       RETURN_IF_NOT_OK(MaskHelper<int64_t>(input, *output, casted_value, op));
       break;
+#ifndef ENABLE_MD_LITE_X86_64
     case DataType::DE_FLOAT16:
       RETURN_IF_NOT_OK(MaskHelper<float16>(input, *output, casted_value, op));
       break;
+#endif
     case DataType::DE_FLOAT32:
       RETURN_IF_NOT_OK(MaskHelper<float>(input, *output, casted_value, op));
       break;
@@ -732,6 +746,7 @@ struct UniqueOpHashMap<float16> {
 };
 
 #else
+#ifndef ENABLE_MD_LITE_X86_64
 struct gn_hash {
   size_t operator()(const float16 &f) const { return static_cast<std::size_t>(f); }
 };
@@ -740,7 +755,7 @@ template <>
 struct UniqueOpHashMap<float16> {
   using map_type = std::unordered_map<float16, int32_t, gn_hash>;
 };
-
+#endif
 #endif
 
 template <>
@@ -809,9 +824,13 @@ Status Unique(const std::shared_ptr<Tensor> &input, std::shared_ptr<Tensor> *out
     RETURN_IF_NOT_OK(UniqueHelper<uint16_t>(input, output, output_idx, output_cnt));
   } else if (input->type() == DataType::DE_UINT8) {
     RETURN_IF_NOT_OK(UniqueHelper<uint8_t>(input, output, output_idx, output_cnt));
-  } else if (input->type() == DataType::DE_FLOAT16) {
+  }
+#ifndef ENABLE_MD_LITE_X86_64
+  else if (input->type() == DataType::DE_FLOAT16) {  // NOLINT
     RETURN_IF_NOT_OK(UniqueHelper<float16>(input, output, output_idx, output_cnt));
-  } else if (input->type() == DataType::DE_FLOAT32) {
+  }
+#endif
+  else if (input->type() == DataType::DE_FLOAT32) {  // NOLINT
     RETURN_IF_NOT_OK(UniqueHelper<float>(input, output, output_idx, output_cnt));
   } else if (input->type() == DataType::DE_FLOAT64) {
     RETURN_IF_NOT_OK(UniqueHelper<double>(input, output, output_idx, output_cnt));

mindspore/core/base/float16.h

@@ -23,6 +23,7 @@
 using float16 = float16_t;
 inline float half_to_float(float16 h) { return static_cast<float>(h); }
 #else
+#ifndef ENABLE_MD_LITE_X86_64
 #include <functional>
 #include "Eigen/Core"
 
@@ -30,4 +31,5 @@ using float16 = Eigen::half;
 using HalfToFloat = std::function<float(float16)>;
 const inline HalfToFloat half_to_float = Eigen::half_impl::half_to_float;
 #endif
+#endif
 #endif  // MINDSPORE_CORE_BASE_FLOAT16_H_

mindspore/lite/CMakeLists.txt

@@ -260,6 +260,9 @@ endif()
 
 if(BUILD_MINDDATA STREQUAL "lite" OR BUILD_MINDDATA STREQUAL "full" OR BUILD_MINDDATA STREQUAL "wrapper")
     add_compile_definitions(ENABLE_ANDROID)
+    if(NOT PLATFORM_ARM32 AND NOT PLATFORM_ARM64)
+        add_compile_definitions(ENABLE_MD_LITE_X86_64)
+    endif()
     add_subdirectory(${CMAKE_CURRENT_SOURCE_DIR}/minddata)
 endif()
 

mindspore/lite/tools/optimizer/common/gllo_utils.cc

@@ -20,11 +20,14 @@
 #include <unordered_map>
 #include <functional>
 #include <string>
+#include "Eigen/Core"
 #include "ops/fusion/conv2d_fusion.h"
 #include "src/common/common.h"
 #include "frontend/operator/ops.h"
 #include "backend/optimizer/common/helper.h"
 
+using float16 = Eigen::half;
+
 namespace mindspore {
 namespace opt {
 namespace {