Add expm1.

theano/scalar/basic.py

@@ -1975,6 +1975,25 @@ class Exp2(UnaryScalarOp):
 exp2 = Exp2(upgrade_to_float, name='exp2')
 
 
+class Expm1(UnaryScalarOp):
+    def impl(self, x):
+        return numpy.expm1(x)
+
+    def grad(self, (x, ), (gz, )):
+        if x.type in complex_types:
+            raise NotImplementedError()
+        elif x.type in float_types:
+            return gz * exp(x),
+        else:
+            return None,
+
+    def c_code(self, node, name, (x, ), (z, ), sub):
+        if node.inputs[0].type in complex_types:
+            raise NotImplementedError('type not supported', type)
+        return "%(z)s = exp(%(x)s) - 1;" % locals()
+expm1 = Expm1(upgrade_to_float, name='expm1')
+
+
 class Sqr(UnaryScalarOp):
     def impl(self, x):
         return x * x

theano/tensor/basic.py

@@ -2578,6 +2578,11 @@ def exp2(a):
     """2^`a`"""
 
 
+@_scal_elemwise_with_nfunc('expm1', 1, 1)
+def expm1(a):
+    """e^`a - 1`"""
+
+
 @_scal_elemwise_with_nfunc('negative', 1, 1)
 def neg(a):
     """-a"""

theano/tensor/inplace.py

@@ -91,6 +91,10 @@ def exp_inplace(a):
 def exp2_inplace(a):
     """2^`a` (inplace on `a`)"""
 
+@_scal_inplace
+def expm1_inplace(a):
+    """e^`a` - 1 (inplace on `a`)"""
+
 @_scal_inplace
 def neg_inplace(a):
     """-a (inplace on a)"""

theano/tensor/tests/test_basic.py

@@ -966,6 +966,18 @@ Exp2InplaceTester = makeBroadcastTester(op=inplace.exp2_inplace,
                                          grad=_grad_broadcast_unary_normal,
                                          inplace=True)
 
+
+Expm1Tester = makeBroadcastTester(op=tensor.expm1,
+                                  expected=numpy.expm1,
+                                  good=_good_broadcast_unary_normal,
+                                  grad=_grad_broadcast_unary_normal)
+Expm1InplaceTester = makeBroadcastTester(op=inplace.expm1_inplace,
+                                         expected=numpy.expm1,
+                                         good=_good_broadcast_unary_normal,
+                                         grad=_grad_broadcast_unary_normal,
+                                         inplace=True)
+
+
 _good_broadcast_unary_positive = dict(normal=(rand_ranged(0.001, 5, (2, 3)),),
                                       integers=(randint_ranged(1, 5, (2, 3)),),
                                       complex=(randc128_ranged(1, 5, (2, 3)),),