diff --git a/impl/ascend/ascend_tensor.hpp b/impl/ascend/ascend_tensor.hpp
index 87b6d4672..46e78ed00 100644
--- a/impl/ascend/ascend_tensor.hpp
+++ b/impl/ascend/ascend_tensor.hpp
@@ -233,7 +233,7 @@ class AscendTensor final {
         return tensor_;
     }
 
-    bool isSame(const AscendTensor& t) const { return this->tensor_ == t.tensor_; }
+    bool isSame(const AscendTensor& t) const { return this->tensorHandle() == t.tensorHandle(); }
 
     int64_t getAclMemBufferSize() const;
     std::vector<int64_t> getAclMemShape() const;
diff --git a/impl/ascend/common/utils.cpp b/impl/ascend/common/utils.cpp
index d3f3e0f59..55e610b19 100644
--- a/impl/ascend/common/utils.cpp
+++ b/impl/ascend/common/utils.cpp
@@ -10,6 +10,7 @@
 #include <cstddef>
 #include <cstdint>
 #include <functional>
+#include <iostream>
 #include <numeric>
 #include <string>
 #include <type_traits>
@@ -182,6 +183,44 @@ diopiError_t reshape(diopiContextHandle_t ctx, const AscendTensor& src, AscendTe
     return diopiSuccess;
 }
 
+AscendTensor reshape(diopiContextHandle_t ctx, const AscendTensor& src, const std::vector<int64_t>& shape) {
+    std::cout << "come into reshape......" << std::endl;
+    ASCEND_CHECK_ABORT(src.defined(), "input tensor is nullptr.");
+
+    if (src.shape() == shape) {
+        std::cout << "shape is same, return src tensor pointer=" << src.tensorHandle() << std::endl;
+        return src;
+    }
+
+    int64_t expectedSize = std::accumulate(shape.begin(), shape.end(), 1, std::multiplies<int64_t>());
+    ASCEND_CHECK_THROW(expectedSize == src.numel(), "reshape size not match, expect %ld, but got %ld", expectedSize, src.numel());
+
+    {
+        diopiTensorHandle_t out = nullptr;
+        diopiSize_t outShape{shape.data(), static_cast<int64_t>(shape.size())};
+        diopiRequireTensor(ctx, &out, &outShape, nullptr, src.dtype(), diopi_device);
+        AscendTensor outAt(out), tmp(src);
+        tmp.view(shape);
+        DIOPI_ASCEND_CALL_ACLNN(aclnnInplaceCopy, ctx, outAt, tmp);
+        return outAt;
+    }
+
+    AscendTensor result;
+    if (src.isContiguous()) {
+        result = src;
+        result.view(shape);
+        std::cout << "isContiguous reshape tensor pointer=" << result.tensorHandle() << std::endl;
+        return result;
+    }
+
+    std::cout << "not isContiguous, make tensor......" << std::endl;
+    makeTensor(ctx, result, shape, src.dtype());
+    std::cout << "make tensor pointer=" << result.tensorHandle() << std::endl;
+    DIOPI_ASCEND_CALL_ACLNN(aclnnInplaceCopy, ctx, result, src);
+
+    return AscendTensor(result.tensorHandle());
+}
+
 diopiError_t aclAsStridedCore(diopiContextHandle_t ctx, const AscendTensor& src, AscendTensor& dst) {
     diopiTensorHandle_t targetObj = const_cast<diopiTensorHandle_t>(static_cast<diopiConstTensorHandle_t>(dst));
     AclOpRunner<4, 1>("AsStrided", ctx)
diff --git a/impl/ascend/common/utils.hpp b/impl/ascend/common/utils.hpp
index 80b1ce056..05314907d 100644
--- a/impl/ascend/common/utils.hpp
+++ b/impl/ascend/common/utils.hpp
@@ -89,6 +89,8 @@ diopiError_t makeTensorFromScalar(diopiContextHandle_t ctx, AscendTensor& dst, c
 
 diopiError_t reshape(diopiContextHandle_t ctx, const AscendTensor& src, AscendTensor& dst, const std::vector<int64_t>& shape);
 
+AscendTensor reshape(diopiContextHandle_t ctx, const AscendTensor& src, const std::vector<int64_t>& shape);
+
 diopiError_t contiguous(diopiContextHandle_t ctx, const AscendTensor& src, AscendTensor& dst, diopiMemoryFormat_t format = diopiMemoryFormat_t::Contiguous);
 
 diopiError_t castTensor(diopiContextHandle_t ctx, const AscendTensor& src, AscendTensor& dst);
diff --git a/impl/ascend/functions/unique.cpp b/impl/ascend/functions/unique.cpp
index dc1e9f9ae..13b80a481 100644
--- a/impl/ascend/functions/unique.cpp
+++ b/impl/ascend/functions/unique.cpp
@@ -62,9 +62,14 @@ diopiError_t diopiUnique(diopiContextHandle_t ctx, diopiTensorHandle_t* out, dio
     ASCEND_CHECK_ABORT(ret == 0, "get out aclGetViewShape failed");
 
     // fill out tensor
-    AscendTensor outReshapeAt;
-    reshape(ctx, outTmpAt, outReshapeAt, {viewDims, viewDims + viewDimNum});
-    *out = const_cast<diopiTensorHandle_t>(outReshapeAt.tensorHandle());
+    if (1) {
+        AscendTensor outReshapeAt;
+        reshape(ctx, outTmpAt, outReshapeAt, {viewDims, viewDims + viewDimNum});
+        *out = const_cast<diopiTensorHandle_t>(outReshapeAt.tensorHandle());
+    } else {
+        AscendTensor outReshapeAt = reshape(ctx, outTmpAt, {viewDims, viewDims + viewDimNum});
+        *out = const_cast<diopiTensorHandle_t>(outReshapeAt.tensorHandle());
+    }
 
     // fill indices tensor
     if (returnInverse) {
@@ -78,9 +83,14 @@ diopiError_t diopiUnique(diopiContextHandle_t ctx, diopiTensorHandle_t* out, dio
         int ret2 = aclGetViewShape(std::get<countsTensorIndex>(params), &viewDims, &viewDimNum);
         ASCEND_CHECK_ABORT(ret2 == 0, "get count aclGetViewShape failed");
 
-        AscendTensor countsReshapeAt;
-        reshape(ctx, countsTmpAt, countsReshapeAt, {viewDims, viewDims + viewDimNum});
-        *counts = const_cast<diopiTensorHandle_t>(countsReshapeAt.tensorHandle());
+        if (1) {
+            AscendTensor countsReshapeAt;
+            reshape(ctx, countsTmpAt, countsReshapeAt, {viewDims, viewDims + viewDimNum});
+            *counts = const_cast<diopiTensorHandle_t>(countsReshapeAt.tensorHandle());
+        } else {
+            AscendTensor countsReshapeAt = reshape(ctx, countsTmpAt, {viewDims, viewDims + viewDimNum});
+            *counts = const_cast<diopiTensorHandle_t>(countsReshapeAt.tensorHandle());
+        }
     }
 
     // delete viewDims pointer