[TOPI] Fix declaration for different dtypes (#546)

python/tvm/expr.py

@@ -18,6 +18,7 @@
 from __future__ import absolute_import as _abs
 from ._ffi.node import NodeBase, register_node
 from . import make as _make
+from . import _api_internal
 
 class ExprOp(object):
     def __add__(self, other):
@@ -60,7 +61,8 @@ def __mod__(self, other):
         return _make.Mod(self, other)
 
     def __neg__(self):
-        return self.__mul__(-1)
+        neg_one = _api_internal._const(-1, self.dtype)
+        return self.__mul__(neg_one)
 
     def __lshift__(self, other):
         return _make.Call(self.dtype, "shift_left", [self, other], Call.PureIntrinsic, None, 0)

topi/python/topi/nn/elemwise.py

@@ -17,7 +17,7 @@ def relu(x):
     y : tvm.Tensor
         The result.
     """
-    return tvm.compute(x.shape, lambda *i: tvm.max(x(*i), 0))
+    return tvm.compute(x.shape, lambda *i: tvm.max(x(*i), tvm.const(0, x.dtype)))
 
 
 @tvm.tag_scope(tag=tag.ELEMWISE)

topi/python/topi/nn/pooling.py

@@ -38,7 +38,7 @@ def global_pool(data, pool_type):
                             tvm.sum(data[n, c, dheight, dwidth], axis=[dheight, dwidth]), \
                             tag="global_pool_sum")
         return tvm.compute((batch, channel, 1, 1), lambda n, c, h, w: \
-                            tsum[n, c, h, w] / (height*width), \
+                            tsum[n, c, h, w] / (height*width).astype(tsum.dtype), \
                             tag=tag.ELEMWISE)
     else:
         raise ValueError("Pool type should be 'avg' or 'max'.")