add unique yaml and final state api (#41460)

python/paddle/fluid/tests/unittests/test_unique.py

@@ -21,6 +21,7 @@ import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.op import Operator
+from paddle.fluid.framework import _test_eager_guard
 
 
 class TestUniqueOp(OpTest):
@@ -251,6 +252,12 @@ class TestUniqueAPI(unittest.TestCase):
         self.assertTrue((counts.numpy() == np_counts).all(), True)
         paddle.enable_static()
 
+    def test_dygraph_final_state_api(self):
+        with _test_eager_guard():
+            self.test_dygraph_api_out()
+            self.test_dygraph_api_attr()
+            self.test_dygraph_attr_dtype()
+
     def test_static_graph(self):
         with paddle.static.program_guard(paddle.static.Program(),
                                          paddle.static.Program()):

python/paddle/tensor/manipulation.py

@@ -1211,11 +1211,16 @@ def unique(x,
     else:
         axis = [axis]
     attr_dtype = convert_np_dtype_to_dtype_(dtype)
-    if paddle.in_dynamic_mode():
+    if _non_static_mode():
+        if in_dygraph_mode():
+            out, indices, inverse, counts = _C_ops.final_state_unique(
+                x, return_index, return_inverse, return_counts, axis,
+                attr_dtype)
+        if _in_legacy_dygraph():
             out, inverse, indices, counts = _C_ops.unique(
                 x, 'dtype', attr_dtype, 'return_index', return_index,
-            'return_inverse', return_inverse, 'return_counts', return_counts,
-            'axis', axis, "is_sorted", True)
+                'return_inverse', return_inverse, 'return_counts',
+                return_counts, 'axis', axis, "is_sorted", True)
         outs = [out]
         if return_index:
             outs.append(indices)

python/paddle/utils/code_gen/api.yaml

@@ -1937,6 +1937,16 @@
     func : unfold
   backward : unfold_grad
 
+# The `axis` argument of Python API paddle.unique is not vector
+- api : unique
+  args : (Tensor x, bool return_index, bool return_inverse, bool return_counts, int[] axis, DataType dtype=DataType::INT64)
+  output : Tensor(out), Tensor(indices), Tensor(inverse), Tensor(counts)
+  infer_meta :
+    func : UniqueInferMeta
+  kernel :
+    func : unique
+    data_type : x
+
 - api : unsqueeze
   args : (Tensor x, IntArray axes)
   output : Tensor(xshape), Tensor(out)