[Fixbug] Fix a bug in the minimum and maximum operator (#102)

* fix a typo in min/max; change signature

* add test

python/hidet/graph/ops/definitions/arithmetic.py

@@ -415,7 +415,7 @@ def __init__(self, cond: Tensor, x: Tensor, y: Tensor):
 
 
 class MaxOp(Operator):
-    def __init__(self, tensors: List[Tensor]):
+    def __init__(self, *tensors: Tensor):
         def scalar_max(args: List[expr.Expr]):
             if len(args) == 1:
                 return args[0]
@@ -434,12 +434,12 @@ def scalar_max(args: List[expr.Expr]):
 
 
 class MinOp(Operator):
-    def __init__(self, tensors: List[Tensor]):
+    def __init__(self, *tensors: Tensor):
         def scalar_min(args: List[expr.Expr]):
             if len(args) == 1:
                 return args[0]
             else:
-                return primitives.max(args[0], scalar_min(args[1:]))
+                return primitives.min(args[0], scalar_min(args[1:]))
 
         super().__init__(
             inputs=list(tensors),
@@ -671,12 +671,12 @@ def where(cond: Tensor, x: Tensor, y: Tensor) -> Tensor:
 
 def maximum(a: Tensor, b: Tensor, *others: Tensor) -> Tensor:
     args = [a, b] + list(others)
-    return MaxOp(args).get_output(0)
+    return MaxOp(*args).get_output(0)
 
 
 def minimum(a: Tensor, b: Tensor, *others: Tensor) -> Tensor:
     args = [a, b] + list(others)
-    return MinOp(args).get_output(0)
+    return MinOp(*args).get_output(0)
 
 
 def mod(x: Tensor, y: Tensor) -> Tensor:

tests/graph/operators/test_arithmetic.py

@@ -125,6 +125,16 @@ def test_bitwise_xor(a_shape, b_shape):
     check_binary(a_shape, b_shape, np.int32, np.bitwise_xor, ops.bitwise_xor)
 
 
+@pytest.mark.parametrize("a_shape, b_shape", binary_op_shapes)
+def test_minimum(a_shape, b_shape):
+    check_binary(a_shape, b_shape, np.int32, np.minimum, ops.minimum)
+
+
+@pytest.mark.parametrize("a_shape, b_shape", binary_op_shapes)
+def test_maximum(a_shape, b_shape):
+    check_binary(a_shape, b_shape, np.int32, np.maximum, ops.maximum)
+
+
 @pytest.mark.parametrize("a_shape", unary_op_shapes)
 def test_ceil(a_shape):
     check_unary(a_shape, np.float32, np.ceil, ops.ceil)