Support for sign (#2775)

docs/api/python/topi.rst

@@ -11,6 +11,7 @@ List of operators
    topi.negative
    topi.floor
    topi.ceil
+   topi.sign
    topi.trunc
    topi.round
    topi.abs
@@ -96,6 +97,7 @@ topi
 .. autofunction:: topi.identity
 .. autofunction:: topi.floor
 .. autofunction:: topi.ceil
+.. autofunction:: topi.sign
 .. autofunction:: topi.trunc
 .. autofunction:: topi.round
 .. autofunction:: topi.abs

docs/langref/relay_op.rst

@@ -81,6 +81,7 @@ This level enables additional math and transform operators.
    tvm.relay.squeeze
    tvm.relay.floor
    tvm.relay.ceil
+   tvm.relay.sign
    tvm.relay.trunc
    tvm.relay.clip
    tvm.relay.round
@@ -213,6 +214,7 @@ Level 3 Definitions
 .. autofunction:: tvm.relay.squeeze
 .. autofunction:: tvm.relay.floor
 .. autofunction:: tvm.relay.ceil
+.. autofunction:: tvm.relay.sign
 .. autofunction:: tvm.relay.trunc
 .. autofunction:: tvm.relay.clip
 .. autofunction:: tvm.relay.round

python/tvm/relay/op/_tensor.py

@@ -16,6 +16,7 @@ register_schedule("floor", schedule_broadcast)
 register_schedule("ceil", schedule_broadcast)
 register_schedule("trunc", schedule_broadcast)
 register_schedule("round", schedule_broadcast)
+register_schedule("sign", schedule_broadcast)
 register_schedule("abs", schedule_broadcast)
 register_schedule("tanh", schedule_broadcast)
 register_schedule("logical_not", schedule_broadcast)

python/tvm/relay/op/tensor.py

@@ -158,6 +158,20 @@ def abs(data):
     """
     return _make.abs(data)
 
+def sign(data):
+    """Compute element-wise absolute of data.
+
+    Parameters
+    ----------
+    data : relay.Expr
+        The input data
+
+    Returns
+    -------
+    result : relay.Expr
+        The computed result.
+    """
+    return _make.sign(data)
 
 def tanh(data):
     """Compute element-wise tanh of data.

src/relay/op/tensor/unary.cc

@@ -146,6 +146,16 @@ RELAY_REGISTER_UNARY_OP("round")
 .set_support_level(3)
 .set_attr<FTVMCompute>("FTVMCompute", RELAY_UNARY_COMPUTE(topi::round));
 
+RELAY_REGISTER_UNARY_OP("sign")
+.describe(R"code(Returns the sign of input array, computed element-wise.
+
+.. numpy::
+   sign(x)
+
+)code" TVM_ADD_FILELINE)
+.set_support_level(3)
+.set_attr<FTVMCompute>("FTVMCompute", RELAY_UNARY_COMPUTE(topi::sign));
+
 
 RELAY_REGISTER_UNARY_OP("abs")
 .describe(R"code(Returns the abs of input array, computed element-wise.

tests/python/relay/test_op_level3.py

@@ -25,7 +25,8 @@ def test_unary_identity():
                (relay.round, np.round),
                (relay.abs, np.abs),
                (relay.copy, None), # np.copy
-               (relay.negative, np.negative)]:
+               (relay.negative, np.negative),
+               (relay.sign, np.sign)]:
         shape = (8, 9, 4)
         x = relay.var("x", relay.TensorType(shape, "float32"))
         y = op(x)

topi/include/topi/elemwise.h

@@ -89,6 +89,28 @@ inline Tensor logical_not(const Tensor& x,
 }
 
 /*!
+* \brief Returns the sign of the tensor
+*
+* \param x The input tensor
+* \param name The name of the operation
+* \param tag The tag to mark the operation
+*
+* \return A Tensor whose op member is the sign
+*/
+inline Tensor sign(const Tensor& x,
+                   std::string name = "tensor",
+                   std::string tag = kElementWise) {
+  return compute(x->shape, [&](const Array<Var>& i) {
+    Expr zero = make_zero(x->dtype);
+    Expr one = make_const(x->dtype, 1);
+    Expr minus_one = make_const(x->dtype, -1);
+    auto s1 = tvm::ir::Select::make((x(i) < zero), minus_one, zero);
+    auto s2 = tvm::ir::Select::make((x(i) > zero), one, s1);
+    return s2;
+  }, name, tag);
+}
+
+/*!
 * \brief Creates an operation that clips each element of a tensor to
 * the interval [a_min, a_max]
 *

topi/python/topi/math.py

@@ -3,6 +3,7 @@
 from __future__ import absolute_import as _abs
 import tvm
 from . import tag
+from . import cpp
 
 @tvm.tag_scope(tag=tag.ELEMWISE)
 def identity(x):
@@ -107,6 +108,20 @@ def ceil(x):
     """
     return tvm.compute(x.shape, lambda *i: tvm.ceil(x(*i)))
 
+def sign(x):
+    """Returns -1, 0, 1 based on sign of x.
+
+    Parameters
+    ----------
+    x : tvm.Tensor
+        Input argument.
+
+    Returns
+    -------
+    y : tvm.Tensor
+        The result.
+    """
+    return cpp.sign(x)
 
 @tvm.tag_scope(tag=tag.ELEMWISE)
 def trunc(x):

topi/src/topi.cc

@@ -173,6 +173,11 @@ TVM_REGISTER_GLOBAL("topi.elemwise_sum")
   *rv = elemwise_sum(args[0]);
   });
 
+TVM_REGISTER_GLOBAL("topi.sign")
+.set_body([](TVMArgs args, TVMRetValue *rv) {
+  *rv = sign(args[0]);
+  });
+
 TVM_REGISTER_GLOBAL("topi.full")
 .set_body([](TVMArgs args, TVMRetValue *rv) {
   *rv = full(args[0], args[1], args[2]);

topi/tests/python/test_topi_math.py

@@ -18,9 +18,10 @@ def test_ewise():
 
     shape = (20, 3)
 
-    def test_apply(func, name, f_numpy, low, high, check_round=False):
+    def test_apply(func, name, f_numpy, low, high, check_round=False, skip_name_check=False):
         B = func(A)
         assert tuple(B.shape) == tuple(A.shape)
+        if not skip_name_check:
             assert B.op.body[0].name == name
         a_np = np.random.uniform(low=low, high=high, size=shape).astype(A.dtype) * 10
         # avoid round check too close to boundary
@@ -49,6 +50,7 @@ def test_ewise():
 
     test_apply(topi.floor, "floor", np.floor, -100, 100)
     test_apply(topi.ceil, "ceil", np.ceil, -100, 100)
+    test_apply(topi.sign, "sign", np.sign, -100, 100, skip_name_check=True)
     test_apply(topi.trunc, "trunc", np.trunc, -100, 100)
     test_apply(topi.abs, "fabs", np.abs, -100, 100)
     test_apply(topi.round, "round", np.round, -100, 100, check_round=True)