Skip to content

P
pytensor
提交 3089deda authored 作者: Bart van Merrienboer's avatar Bart van Merrienboer
浏览文件
操作

Fix E261: at least two spaces before inline comment

上级 b34bc31d
隐藏空白字符变更
内嵌 并排
正在显示 包含 390 行增加 和 390 行删除
theano/compat/python2x.py
浏览文件 @ 3089deda
...@@ -405,7 +405,7 @@ if sys.version_info[:2] < (2, 7): ...@@ -405,7 +405,7 @@ if sys.version_info[:2] < (2, 7):
for elem, count in iterable.iteritems(): for elem, count in iterable.iteritems():
self[elem] = self_get(elem, 0) + count self[elem] = self_get(elem, 0) + count
else: else:
dict.update(self, iterable) # fast path when counter is empty dict.update(self, iterable) # fast path when counter is empty
else: else:
self_get = self.get self_get = self.get
for elem in iterable: for elem in iterable:
......
theano/compile/debugmode.py
浏览文件 @ 3089deda
...@@ -326,7 +326,7 @@ class BadDestroyMap(DebugModeError): ...@@ -326,7 +326,7 @@ class BadDestroyMap(DebugModeError):
wasn't in the destroy_map""" wasn't in the destroy_map"""
def __init__(self, node, idx, old_val, new_val, perform): def __init__(self, node, idx, old_val, new_val, perform):
#super(BadDestroyMap, self).__init__() #super(BadDestroyMap, self).__init__()
DebugModeError.__init__(self)#to be compatible with python2.4 DebugModeError.__init__(self) # to be compatible with python2.4
self.node = node self.node = node
self.idx = idx self.idx = idx
self.old_val = old_val self.old_val = old_val
...@@ -794,7 +794,7 @@ def _check_inputs(node, storage_map, r_vals, dr_vals, active_nodes, ...@@ -794,7 +794,7 @@ def _check_inputs(node, storage_map, r_vals, dr_vals, active_nodes,
# bad: there should only be one active node that destroys any variable # bad: there should only be one active node that destroys any variable
raise Exception('failure in topological ordering') raise Exception('failure in topological ordering')
if clobber_dr_vals: if clobber_dr_vals:
dr_vals[r] = (storage_map[r][0], node) #no copy, this is the last use of this variable dr_vals[r] = (storage_map[r][0], node) # no copy, this is the last use of this variable
# make sure that dr_vals[r] doens't get used again # make sure that dr_vals[r] doens't get used again
storage_map[r][0] = data_destroyed storage_map[r][0] = data_destroyed
else: else:
...@@ -887,7 +887,7 @@ def _check_strides_match(a, b, warn_err, op): ...@@ -887,7 +887,7 @@ def _check_strides_match(a, b, warn_err, op):
try: try:
strides_eq = a.strides == b.strides strides_eq = a.strides == b.strides
except Exception: except Exception:
return # no strides return # no strides
if not strides_eq: if not strides_eq:
e = TypeError('Stride mismatch', (a.shape, b.shape, a.strides, e = TypeError('Stride mismatch', (a.shape, b.shape, a.strides,
...@@ -1643,7 +1643,7 @@ class _Linker(gof.link.LocalLinker): ...@@ -1643,7 +1643,7 @@ class _Linker(gof.link.LocalLinker):
order_outputs.reverse() order_outputs.reverse()
order = graph.io_toposort(fgraph.inputs, order_outputs) order = graph.io_toposort(fgraph.inputs, order_outputs)
active_order = self.schedule(fgraph) # an ordering of just the active nodes active_order = self.schedule(fgraph) # an ordering of just the active nodes
active_order_set = set(active_order) active_order_set = set(active_order)
# Disable no_recycling, in order to be able to use # Disable no_recycling, in order to be able to use
...@@ -1999,7 +1999,7 @@ class _Linker(gof.link.LocalLinker): ...@@ -1999,7 +1999,7 @@ class _Linker(gof.link.LocalLinker):
raise InvalidValueError(r, storage_map[r][0], hint='c output') raise InvalidValueError(r, storage_map[r][0], hint='c output')
if thunk_py: if thunk_py:
assert r in r_vals #because we put it in during the thunk_py branch assert r in r_vals # because we put it in during the thunk_py branch
# check for stride correctness (may raise exception) # check for stride correctness (may raise exception)
_check_strides_match(r_vals[r], _check_strides_match(r_vals[r],
storage_map[r][0], storage_map[r][0],
...@@ -2032,7 +2032,7 @@ class _Linker(gof.link.LocalLinker): ...@@ -2032,7 +2032,7 @@ class _Linker(gof.link.LocalLinker):
#print >> sys.stderr, i, "DEBUGMODE storing reference output %x" % id(storage_map[r][0]) #print >> sys.stderr, i, "DEBUGMODE storing reference output %x" % id(storage_map[r][0])
# retrieve each output from the storage_map # retrieve each output from the storage_map
r_vals[r] = storage_map[r][0] r_vals[r] = storage_map[r][0]
storage_map[r][0] = None #clear the storage_map for the thunk_c storage_map[r][0] = None # clear the storage_map for the thunk_c
if self.maker.mode.check_preallocated_output: if self.maker.mode.check_preallocated_output:
prealloc_modes = \ prealloc_modes = \
......
theano/compile/function_module.py
浏览文件 @ 3089deda
...@@ -176,7 +176,7 @@ class AliasedMemoryError(Exception): ...@@ -176,7 +176,7 @@ class AliasedMemoryError(Exception):
### ###
DUPLICATE = ['DUPLICATE'] # unique id object used as a placeholder for duplicate entries DUPLICATE = ['DUPLICATE'] # unique id object used as a placeholder for duplicate entries
class Function(object): class Function(object):
""" """
Type of the functions returned by theano.function or theano.FunctionMaker.create. Type of the functions returned by theano.function or theano.FunctionMaker.create.
...@@ -320,7 +320,7 @@ class Function(object): ...@@ -320,7 +320,7 @@ class Function(object):
# Initialize the storage # Initialize the storage
# this loop works by modifying the elements (as variable c) of self.input_storage inplace. # this loop works by modifying the elements (as variable c) of self.input_storage inplace.
for i, ((input, indices, sinputs), (required, refeed, value)) in enumerate(zip(self.indices, defaults)): for i, ((input, indices, sinputs), (required, refeed, value)) in enumerate(zip(self.indices, defaults)):
if indices is None: # this is true iff input is not a SymbolicInputKit if indices is None: # this is true iff input is not a SymbolicInputKit
c = containers[0] #containers is being used as a stack. Here we pop off the next one. c = containers[0] #containers is being used as a stack. Here we pop off the next one.
c.strict = getattr(input, 'strict', False) c.strict = getattr(input, 'strict', False)
c.allow_downcast = getattr(input, 'allow_downcast', None) c.allow_downcast = getattr(input, 'allow_downcast', None)
...@@ -338,7 +338,7 @@ class Function(object): ...@@ -338,7 +338,7 @@ class Function(object):
c.value = value c.value = value
c.required = required c.required = required
c.implicit = input.implicit c.implicit = input.implicit
c.provided = 0 # this is a count of how many times the input has been provided (reinitialized to 0 on __call__) c.provided = 0 # this is a count of how many times the input has been provided (reinitialized to 0 on __call__)
finder[i] = c finder[i] = c
finder[input.variable] = c finder[input.variable] = c
if input.name not in finder: if input.name not in finder:
...@@ -726,7 +726,7 @@ def _pickle_Function(f): ...@@ -726,7 +726,7 @@ def _pickle_Function(f):
# HACK to detect aliased storage. # HACK to detect aliased storage.
# This is here because aliased relationships are not [currently] preserved across the pickle operation # This is here because aliased relationships are not [currently] preserved across the pickle operation
if not (f.pickle_aliased_memory_strategy == 'ignore'): if not (f.pickle_aliased_memory_strategy == 'ignore'):
all_data = input_storage + inputs_data # addition here means list append all_data = input_storage + inputs_data # addition here means list append
for i, d_i in enumerate(all_data): for i, d_i in enumerate(all_data):
for j, d_j in enumerate(all_data): for j, d_j in enumerate(all_data):
if (i < j) and isinstance(d_i, numpy.ndarray) and isinstance(d_j, numpy.ndarray): if (i < j) and isinstance(d_i, numpy.ndarray) and isinstance(d_j, numpy.ndarray):
......
theano/compile/profilemode.py
浏览文件 @ 3089deda
...@@ -402,7 +402,7 @@ class ProfileMode(Mode): ...@@ -402,7 +402,7 @@ class ProfileMode(Mode):
sop_time={} sop_time={}
sop_call={} sop_call={}
sop_op = {} sop_op = {}
sop_cimpl={} #map each op class to Bool. True iff all applies were done in c. sop_cimpl={} # map each op class to Bool. True iff all applies were done in c.
for a, t in op_time.items(): for a, t in op_time.items():
typ = type(a) typ = type(a)
sop_time.setdefault(typ, 0) sop_time.setdefault(typ, 0)
......
theano/compile/tests/test_function_module.py
浏览文件 @ 3089deda
...@@ -36,7 +36,7 @@ def checkfor(testcase, fn, E): ...@@ -36,7 +36,7 @@ def checkfor(testcase, fn, E):
class T_function(unittest.TestCase): class T_function(unittest.TestCase):
def test_none(self): def test_none(self):
fn = function([], None) #ok fn = function([], None) # ok
rval = fn() rval = fn()
if rval == []: if rval == []:
raise KnownFailureTest('See #254: Using None as function output leads to [] return value') raise KnownFailureTest('See #254: Using None as function output leads to [] return value')
...@@ -44,7 +44,7 @@ class T_function(unittest.TestCase): ...@@ -44,7 +44,7 @@ class T_function(unittest.TestCase):
assert rval is None assert rval is None
def test_empty(self): def test_empty(self):
fn = function([], []) #ok fn = function([], []) # ok
self.assertTrue(fn() == []) self.assertTrue(fn() == [])
def test_extra_inputs(self): def test_extra_inputs(self):
...@@ -122,22 +122,22 @@ class T_function(unittest.TestCase): ...@@ -122,22 +122,22 @@ class T_function(unittest.TestCase):
self.assertTrue(f(s=2, x=1) == 0.5) self.assertTrue(f(s=2, x=1) == 0.5)
self.assertTrue(f(x=2, s=1) == 2.0) self.assertTrue(f(x=2, s=1) == 2.0)
self.assertTrue(f(2, s=1) == 2.0) self.assertTrue(f(2, s=1) == 2.0)
checkfor(self, lambda : f(2, x=2.0), TypeError) #got multiple values for keyword argument 'x' checkfor(self, lambda : f(2, x=2.0), TypeError) # got multiple values for keyword argument 'x'
checkfor(self, lambda : f(x=1), TypeError) #takes exactly 2 non-keyword arguments (1 given) checkfor(self, lambda : f(x=1), TypeError) # takes exactly 2 non-keyword arguments (1 given)
checkfor(self, lambda : f(s=1), TypeError) #takes exactly 2 non-keyword arguments (0 given) checkfor(self, lambda : f(s=1), TypeError) # takes exactly 2 non-keyword arguments (0 given)
def test_naming_rule1(self): def test_naming_rule1(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([a, s], a/s) f = function([a, s], a/s)
self.assertTrue(f(1, 2) == 0.5) self.assertTrue(f(1, 2) == 0.5)
self.assertTrue(f(2, 1) == 2.0) self.assertTrue(f(2, 1) == 2.0)
self.assertTrue(f(2, s=1) == 2.0) self.assertTrue(f(2, s=1) == 2.0)
checkfor(self, lambda: f(q=2, s=1), TypeError) #got unexpected keyword argument 'q' checkfor(self, lambda: f(q=2, s=1), TypeError) # got unexpected keyword argument 'q'
checkfor(self, lambda: f(a=2, s=1), TypeError) #got unexpected keyword argument 'a' checkfor(self, lambda: f(a=2, s=1), TypeError) # got unexpected keyword argument 'a'
def test_naming_rule2(self): def test_naming_rule2(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
# x's name is ignored because it is followed by anonymous parameter a. # x's name is ignored because it is followed by anonymous parameter a.
...@@ -146,40 +146,40 @@ class T_function(unittest.TestCase): ...@@ -146,40 +146,40 @@ class T_function(unittest.TestCase):
self.assertTrue(f(9, 1, 2) == 0.5) self.assertTrue(f(9, 1, 2) == 0.5)
self.assertTrue(f(9, 2, 1) == 2.0) self.assertTrue(f(9, 2, 1) == 2.0)
self.assertTrue(f(9, 2, s=1) == 2.0) self.assertTrue(f(9, 2, s=1) == 2.0)
checkfor(self, lambda: f(x=9, a=2, s=1), TypeError) #got unexpected keyword argument 'x' checkfor(self, lambda: f(x=9, a=2, s=1), TypeError) # got unexpected keyword argument 'x'
checkfor(self, lambda: f(5.0, x=9), TypeError) #got unexpected keyword argument 'x' checkfor(self, lambda: f(5.0, x=9), TypeError) # got unexpected keyword argument 'x'
def test_naming_rule3(self): def test_naming_rule3(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
# x's name is not ignored (as in test_naming_rule2) because a has a default value. # x's name is not ignored (as in test_naming_rule2) because a has a default value.
f = function([x, In(a, value=1.0), s], a/s+x) f = function([x, In(a, value=1.0), s], a/s+x)
self.assertTrue(f(9, 2, 4) == 9.5) #can specify all args in order self.assertTrue(f(9, 2, 4) == 9.5) # can specify all args in order
self.assertTrue(f(9, 2, s=4) == 9.5) # can give s as kwarg self.assertTrue(f(9, 2, s=4) == 9.5) # can give s as kwarg
self.assertTrue(f(9, s=4) == 9.25) # can give s as kwarg, get default a self.assertTrue(f(9, s=4) == 9.25) # can give s as kwarg, get default a
self.assertTrue(f(x=9, s=4) == 9.25) # can give s as kwarg, omit a, x as kw self.assertTrue(f(x=9, s=4) == 9.25) # can give s as kwarg, omit a, x as kw
checkfor(self, lambda: f(x=9, a=2, s=4), TypeError) #got unexpected keyword argument 'a' checkfor(self, lambda: f(x=9, a=2, s=4), TypeError) # got unexpected keyword argument 'a'
checkfor(self, lambda: f(), TypeError) #takes exactly 3 non-keyword arguments (0 given) checkfor(self, lambda: f(), TypeError) # takes exactly 3 non-keyword arguments (0 given)
checkfor(self, lambda: f(x=9), TypeError) #takes exactly 3 non-keyword arguments (1 given) checkfor(self, lambda: f(x=9), TypeError) # takes exactly 3 non-keyword arguments (1 given)
def test_naming_rule4(self): def test_naming_rule4(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), s], a/s+x) f = function([x, In(a, value=1.0, name='a'), s], a/s+x)
self.assertTrue(f(9, 2, 4) == 9.5) #can specify all args in order self.assertTrue(f(9, 2, 4) == 9.5) # can specify all args in order
self.assertTrue(f(9, 2, s=4) == 9.5) # can give s as kwarg self.assertTrue(f(9, 2, s=4) == 9.5) # can give s as kwarg
self.assertTrue(f(9, s=4) == 9.25) # can give s as kwarg, get default a self.assertTrue(f(9, s=4) == 9.25) # can give s as kwarg, get default a
self.assertTrue(f(9, a=2, s=4) == 9.5) # can give s as kwarg, a as kwarg self.assertTrue(f(9, a=2, s=4) == 9.5) # can give s as kwarg, a as kwarg
self.assertTrue(f(x=9, a=2, s=4) == 9.5) # can give all kwargs self.assertTrue(f(x=9, a=2, s=4) == 9.5) # can give all kwargs
self.assertTrue(f(x=9, s=4) == 9.25) # can give all kwargs self.assertTrue(f(x=9, s=4) == 9.25) # can give all kwargs
checkfor(self, lambda: f(), TypeError) #takes exactly 3 non-keyword arguments (0 given) checkfor(self, lambda: f(), TypeError) # takes exactly 3 non-keyword arguments (0 given)
checkfor(self, lambda: f(5.0, x=9), TypeError) #got multiple values for keyword argument 'x' checkfor(self, lambda: f(5.0, x=9), TypeError) # got multiple values for keyword argument 'x'
def test_state_access(self): def test_state_access(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x)], s+a*x)
...@@ -188,14 +188,14 @@ class T_function(unittest.TestCase): ...@@ -188,14 +188,14 @@ class T_function(unittest.TestCase):
self.assertTrue(f[s] == 0.0) self.assertTrue(f[s] == 0.0)
self.assertTrue(f(3.0) == 3.0) self.assertTrue(f(3.0) == 3.0)
self.assertTrue(f(3.0, a=2.0) == 9.0) #3.0 + 2*3.0 self.assertTrue(f(3.0, a=2.0) == 9.0) # 3.0 + 2*3.0
self.assertTrue(f[a] == 1.0) #state hasn't changed permanently, we just overrode it last line self.assertTrue(f[a] == 1.0) # state hasn't changed permanently, we just overrode it last line
self.assertTrue(f[s] == 9.0) self.assertTrue(f[s] == 9.0)
f[a] = 5.0 f[a] = 5.0
self.assertTrue(f[a] == 5.0) self.assertTrue(f[a] == 5.0)
self.assertTrue(f(3.0) == 24.0) #9 + 3*5 self.assertTrue(f(3.0) == 24.0) # 9 + 3*5
self.assertTrue(f[s] == 24.0) self.assertTrue(f[s] == 24.0)
def test_same_names(self): def test_same_names(self):
...@@ -217,7 +217,7 @@ class T_function(unittest.TestCase): ...@@ -217,7 +217,7 @@ class T_function(unittest.TestCase):
checkfor(self, t, TypeError) checkfor(self, t, TypeError)
def test_copy(self): def test_copy(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x)
...@@ -229,17 +229,17 @@ class T_function(unittest.TestCase): ...@@ -229,17 +229,17 @@ class T_function(unittest.TestCase):
self.assertFalse(g.container[a].storage is f.container[a].storage) self.assertFalse(g.container[a].storage is f.container[a].storage)
self.assertFalse(g.container[s].storage is f.container[s].storage) self.assertFalse(g.container[s].storage is f.container[s].storage)
self.assertFalse(g.value[a] is not f.value[a]) # should not have been copied self.assertFalse(g.value[a] is not f.value[a]) # should not have been copied
self.assertFalse(g.value[s] is f.value[s]) # should have been copied because it is mutable. self.assertFalse(g.value[s] is f.value[s]) # should have been copied because it is mutable.
self.assertFalse((g.value[s] != f.value[s]).any()) # its contents should be identical self.assertFalse((g.value[s] != f.value[s]).any()) # its contents should be identical
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
f(1, 2) # put them out of sync f(1, 2) # put them out of sync
self.assertFalse(f(1, 2) == g(1, 2)) #they should not be equal anymore. self.assertFalse(f(1, 2) == g(1, 2)) # they should not be equal anymore.
def test_shared_state0(self): def test_shared_state0(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x)
...@@ -253,7 +253,7 @@ class T_function(unittest.TestCase): ...@@ -253,7 +253,7 @@ class T_function(unittest.TestCase):
self.assertTrue(g[s] == 0) self.assertTrue(g[s] == 0)
def test_shared_state1(self): def test_shared_state1(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x)
...@@ -268,7 +268,7 @@ class T_function(unittest.TestCase): ...@@ -268,7 +268,7 @@ class T_function(unittest.TestCase):
self.assertTrue(g[s] == 4) self.assertTrue(g[s] == 4)
def test_shared_state2(self): def test_shared_state2(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x,
...@@ -281,7 +281,7 @@ class T_function(unittest.TestCase): ...@@ -281,7 +281,7 @@ class T_function(unittest.TestCase):
f(1, 2) f(1, 2)
self.assertTrue(f[s] == 4) self.assertTrue(f[s] == 4)
self.assertTrue(g[s] == 4) self.assertTrue(g[s] == 4)
g(1, 2) # has no effect on state g(1, 2) # has no effect on state
self.assertTrue(f[s] == 4) self.assertTrue(f[s] == 4)
self.assertTrue(g[s] == 4) self.assertTrue(g[s] == 4)
...@@ -357,20 +357,20 @@ class T_function(unittest.TestCase): ...@@ -357,20 +357,20 @@ class T_function(unittest.TestCase):
a = T.dmatrix() a = T.dmatrix()
f = function([a], Out(a, borrow=False)) f = function([a], Out(a, borrow=False))
o = N.ones((3, 3)) o = N.ones((3, 3))
assert o is not f(o) #function no longer permits aliasing outputs to inputs assert o is not f(o) # function no longer permits aliasing outputs to inputs
f = function([a], Out(a*4, borrow=False)) f = function([a], Out(a*4, borrow=False))
o = N.ones((3, 3)) o = N.ones((3, 3))
four = f(o) four = f(o)
assert numpy.all(four==4) assert numpy.all(four==4)
f(o+.1) #should not clobber the memory used to store four f(o+.1) # should not clobber the memory used to store four
assert numpy.all(four==4) assert numpy.all(four==4)
f = function([a], Out(a*4, borrow=True), mode=theano.Mode('c|py_nogc', 'fast_run')) f = function([a], Out(a*4, borrow=True), mode=theano.Mode('c|py_nogc', 'fast_run'))
o = N.ones((3, 3)) o = N.ones((3, 3))
four = f(o) four = f(o)
assert numpy.all(four==4) assert numpy.all(four==4)
f(o+.1) #should clobber the memory used to store four f(o+.1) # should clobber the memory used to store four
if theano.config.cxx: if theano.config.cxx:
assert not numpy.all(four==4) assert not numpy.all(four==4)
else: else:
...@@ -423,7 +423,7 @@ class T_function(unittest.TestCase): ...@@ -423,7 +423,7 @@ class T_function(unittest.TestCase):
class T_picklefunction(unittest.TestCase): class T_picklefunction(unittest.TestCase):
def test_deepcopy(self): def test_deepcopy(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x)
...@@ -452,16 +452,16 @@ class T_picklefunction(unittest.TestCase): ...@@ -452,16 +452,16 @@ class T_picklefunction(unittest.TestCase):
for ((f_req, f_feed, f_val), (g_req, g_feed, g_val)) in zip( for ((f_req, f_feed, f_val), (g_req, g_feed, g_val)) in zip(
f.defaults, g.defaults)])) f.defaults, g.defaults)]))
self.assertFalse(g.value[1] is f.value[1]) # should not have been copied self.assertFalse(g.value[1] is f.value[1]) # should not have been copied
self.assertFalse(g.value[2] is f.value[2]) # should have been copied because it is mutable. self.assertFalse(g.value[2] is f.value[2]) # should have been copied because it is mutable.
self.assertFalse((g.value[2] != f.value[2]).any()) # its contents should be identical self.assertFalse((g.value[2] != f.value[2]).any()) # its contents should be identical
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
f(1, 2) # put them out of sync f(1, 2) # put them out of sync
self.assertFalse(f(1, 2) == g(1, 2)) #they should not be equal anymore. self.assertFalse(f(1, 2) == g(1, 2)) # they should not be equal anymore.
g(1, 2) # put them back in sync g(1, 2) # put them back in sync
self.assertTrue(f(3) == g(3)) # They should be in sync again. self.assertTrue(f(3) == g(3)) # They should be in sync again.
def test_deepcopy_shared_container(self): def test_deepcopy_shared_container(self):
# Ensure that shared containers remain shared after a deep copy. # Ensure that shared containers remain shared after a deep copy.
...@@ -488,7 +488,7 @@ class T_picklefunction(unittest.TestCase): ...@@ -488,7 +488,7 @@ class T_picklefunction(unittest.TestCase):
self.assertTrue(fc[ac] == 2) self.assertTrue(fc[ac] == 2)
def test_pickle(self): def test_pickle(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x) f = function([x, In(a, value=1.0, name='a'), In(s, value=0.0, update=s+a*x, mutable=True)], s+a*x)
...@@ -513,17 +513,17 @@ class T_picklefunction(unittest.TestCase): ...@@ -513,17 +513,17 @@ class T_picklefunction(unittest.TestCase):
self.assertFalse(x in g.container) self.assertFalse(x in g.container)
self.assertFalse(x in g.value) self.assertFalse(x in g.value)
self.assertFalse(g.value[1] is f.value[1]) # should not have been copied self.assertFalse(g.value[1] is f.value[1]) # should not have been copied
self.assertFalse(g.value[2] is f.value[2]) # should have been copied because it is mutable. self.assertFalse(g.value[2] is f.value[2]) # should have been copied because it is mutable.
self.assertFalse((g.value[2] != f.value[2]).any()) # its contents should be identical self.assertFalse((g.value[2] != f.value[2]).any()) # its contents should be identical
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
self.assertTrue(f(2, 1) == g(2)) #they should be in sync, default value should be copied. self.assertTrue(f(2, 1) == g(2)) # they should be in sync, default value should be copied.
f(1, 2) # put them out of sync f(1, 2) # put them out of sync
self.assertFalse(f(1, 2) == g(1, 2)) #they should not be equal anymore. self.assertFalse(f(1, 2) == g(1, 2)) # they should not be equal anymore.
def test_optimizations_preserved(self): def test_optimizations_preserved(self):
a = T.dvector() # the a is for 'anonymous' (un-named). a = T.dvector() # the a is for 'anonymous' (un-named).
x = T.dvector('x') x = T.dvector('x')
s = T.dvector('s') s = T.dvector('s')
xm = T.dmatrix('x') xm = T.dmatrix('x')
...@@ -567,7 +567,7 @@ class T_picklefunction(unittest.TestCase): ...@@ -567,7 +567,7 @@ class T_picklefunction(unittest.TestCase):
def test_multiple_functions(self): def test_multiple_functions(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
v = T.vector('v') v = T.vector('v')
...@@ -698,7 +698,7 @@ class T_picklefunction(unittest.TestCase): ...@@ -698,7 +698,7 @@ class T_picklefunction(unittest.TestCase):
class SomethingToPickle(object): class SomethingToPickle(object):
def __init__(self): def __init__(self):
a = T.scalar() # the a is for 'anonymous' (un-named). a = T.scalar() # the a is for 'anonymous' (un-named).
x, s = T.scalars('xs') x, s = T.scalars('xs')
v = T.vector('v') v = T.vector('v')
......
theano/gof/callcache.py
浏览文件 @ 3089deda
...@@ -7,7 +7,7 @@ class CallCache(object): ...@@ -7,7 +7,7 @@ class CallCache(object):
self.filename = filename self.filename = filename
try: try:
if filename is None: if filename is None:
raise IOError('bad filename') #just goes to except raise IOError('bad filename') # just goes to except
f = open(filename, 'r') f = open(filename, 'r')
self.cache = cPickle.load(f) self.cache = cPickle.load(f)
f.close() f.close()
......
theano/gof/destroyhandler.py
浏览文件 @ 3089deda
...@@ -325,11 +325,11 @@ if 0: ...@@ -325,11 +325,11 @@ if 0:
fgraph.destroy_handler = self fgraph.destroy_handler = self
self.fgraph = fgraph self.fgraph = fgraph
self.destroyers = OrderedSet() #set of Apply instances with non-null destroy_map self.destroyers = OrderedSet() # set of Apply instances with non-null destroy_map
self.view_i = {} # variable -> variable used in calculation self.view_i = {} # variable -> variable used in calculation
self.view_o = {} # variable -> set of variables that use this one as a direct input self.view_o = {} # variable -> set of variables that use this one as a direct input
# clients: how many times does an apply use a given variable # clients: how many times does an apply use a given variable
self.clients = {} # variable -> apply -> ninputs self.clients = {} # variable -> apply -> ninputs
self.stale_droot = True self.stale_droot = True
# IG: It's unclear if this is meant to be included in deployed code. It looks like # IG: It's unclear if this is meant to be included in deployed code. It looks like
...@@ -725,11 +725,11 @@ class DestroyHandler(toolbox.Bookkeeper): ...@@ -725,11 +725,11 @@ class DestroyHandler(toolbox.Bookkeeper):
fgraph.destroy_handler = self fgraph.destroy_handler = self
self.fgraph = fgraph self.fgraph = fgraph
self.destroyers = OrderedSet() #set of Apply instances with non-null destroy_map self.destroyers = OrderedSet() # set of Apply instances with non-null destroy_map
self.view_i = OrderedDict() # variable -> variable used in calculation self.view_i = OrderedDict() # variable -> variable used in calculation
self.view_o = OrderedDict() # variable -> set of variables that use this one as a direct input self.view_o = OrderedDict() # variable -> set of variables that use this one as a direct input
# clients: how many times does an apply use a given variable # clients: how many times does an apply use a given variable
self.clients = OrderedDict() # variable -> apply -> ninputs self.clients = OrderedDict() # variable -> apply -> ninputs
self.stale_droot = True self.stale_droot = True
self.debug_all_apps = OrderedSet() self.debug_all_apps = OrderedSet()
...@@ -754,7 +754,7 @@ class DestroyHandler(toolbox.Bookkeeper): ...@@ -754,7 +754,7 @@ class DestroyHandler(toolbox.Bookkeeper):
if self.stale_droot: if self.stale_droot:
droot = OrderedDict() # destroyed view + nonview variables -> foundation droot = OrderedDict() # destroyed view + nonview variables -> foundation
impact = OrderedDict() # destroyed nonview variable -> it + all views of it impact = OrderedDict() # destroyed nonview variable -> it + all views of it
root_destroyer = OrderedDict() # root -> destroyer apply root_destroyer = OrderedDict() # root -> destroyer apply
for app in self.destroyers: for app in self.destroyers:
for output_idx, input_idx_list in app.op.destroy_map.items(): for output_idx, input_idx_list in app.op.destroy_map.items():
......
theano/gof/fg.py
浏览文件 @ 3089deda
...@@ -396,7 +396,7 @@ class FunctionGraph(utils.object2): ...@@ -396,7 +396,7 @@ class FunctionGraph(utils.object2):
# then __prune__ is a no-op. # then __prune__ is a no-op.
for output in node.outputs: for output in node.outputs:
# Cannot prune an op which is an output or used somewhere # Cannot prune an op which is an output or used somewhere
if self.clients(output) or output in self.outputs: #output in self.outputs or self.clients(output): if self.clients(output) or output in self.outputs: # output in self.outputs or self.clients(output):
return return
self.apply_nodes.remove(node) self.apply_nodes.remove(node)
self.variables.difference_update(node.outputs) self.variables.difference_update(node.outputs)
......
theano/gof/link.py
浏览文件 @ 3089deda
...@@ -685,7 +685,7 @@ class PerformLinker(LocalLinker): ...@@ -685,7 +685,7 @@ class PerformLinker(LocalLinker):
f = streamline(fgraph, thunks, order, post_thunk_old_storage, f = streamline(fgraph, thunks, order, post_thunk_old_storage,
no_recycling=no_recycling) no_recycling=no_recycling)
f.allow_gc = self.allow_gc #HACK: this is a way of passing an arg to Function.__call__ f.allow_gc = self.allow_gc # HACK: this is a way of passing an arg to Function.__call__
add_clear_storage(f, computed, storage_map) add_clear_storage(f, computed, storage_map)
f.storage_map = storage_map f.storage_map = storage_map
......
theano/gof/tests/test_graph.py
浏览文件 @ 3089deda
...@@ -142,10 +142,10 @@ class TestClone(X): ...@@ -142,10 +142,10 @@ class TestClone(X):
node = MyOp.make_node(r1, r2) node = MyOp.make_node(r1, r2)
node2 = MyOp.make_node(node.outputs[0], r5) node2 = MyOp.make_node(node.outputs[0], r5)
_, new = clone([r1, r2, r5], node2.outputs, False) _, new = clone([r1, r2, r5], node2.outputs, False)
assert node2.outputs[0].type == new[0].type and node2.outputs[0] is not new[0] # the new output is like the old one but not the same object assert node2.outputs[0].type == new[0].type and node2.outputs[0] is not new[0] # the new output is like the old one but not the same object
assert node2 is not new[0].owner # the new output has a new owner assert node2 is not new[0].owner # the new output has a new owner
assert new[0].owner.inputs[1] is r5 # the inputs are not copied assert new[0].owner.inputs[1] is r5 # the inputs are not copied
assert new[0].owner.inputs[0].type == node.outputs[0].type and new[0].owner.inputs[0] is not node.outputs[0] # check that we copied deeper too assert new[0].owner.inputs[0].type == node.outputs[0].type and new[0].owner.inputs[0] is not node.outputs[0] # check that we copied deeper too
def test_not_destructive(self): def test_not_destructive(self):
# Checks that manipulating a cloned graph leaves the original unchanged. # Checks that manipulating a cloned graph leaves the original unchanged.
......
theano/gof/tests/test_opt.py
浏览文件 @ 3089deda
...@@ -109,18 +109,18 @@ class TestPatternOptimizer: ...@@ -109,18 +109,18 @@ class TestPatternOptimizer:
def test_unification_1(self): def test_unification_1(self):
x, y, z = inputs() x, y, z = inputs()
e = op1(op2(x, x), z) # the arguments to op2 are the same e = op1(op2(x, x), z) # the arguments to op2 are the same
g = Env([x, y, z], [e]) g = Env([x, y, z], [e])
PatternOptimizer((op1, (op2, '1', '1'), '2'), # they are the same in the pattern PatternOptimizer((op1, (op2, '1', '1'), '2'), # they are the same in the pattern
(op4, '2', '1')).optimize(g) (op4, '2', '1')).optimize(g)
# So the replacement should occur # So the replacement should occur
assert str(g) == "[Op4(z, x)]" assert str(g) == "[Op4(z, x)]"
def test_unification_2(self): def test_unification_2(self):
x, y, z = inputs() x, y, z = inputs()
e = op1(op2(x, y), z) # the arguments to op2 are different e = op1(op2(x, y), z) # the arguments to op2 are different
g = Env([x, y, z], [e]) g = Env([x, y, z], [e])
PatternOptimizer((op1, (op2, '1', '1'), '2'), # they are the same in the pattern PatternOptimizer((op1, (op2, '1', '1'), '2'), # they are the same in the pattern
(op4, '2', '1')).optimize(g) (op4, '2', '1')).optimize(g)
# The replacement should NOT occur # The replacement should NOT occur
assert str(g) == "[Op1(Op2(x, y), z)]" assert str(g) == "[Op1(Op2(x, y), z)]"
...@@ -403,7 +403,7 @@ class TestEquilibrium(object): ...@@ -403,7 +403,7 @@ class TestEquilibrium(object):
PatternSub((op4, 'x', 'y'), (op1, 'x', 'y')), PatternSub((op4, 'x', 'y'), (op1, 'x', 'y')),
PatternSub((op3, (op2, 'x', 'y')), (op4, 'x', 'y')) PatternSub((op3, (op2, 'x', 'y')), (op4, 'x', 'y'))
], ],
max_use_ratio=1. / len(g.apply_nodes)) # each opt can only be applied once max_use_ratio=1. / len(g.apply_nodes)) # each opt can only be applied once
opt.optimize(g) opt.optimize(g)
finally: finally:
_logger.setLevel(oldlevel) _logger.setLevel(oldlevel)
......
theano/gof/type.py
浏览文件 @ 3089deda
...@@ -226,8 +226,8 @@ class PureType(object): ...@@ -226,8 +226,8 @@ class PureType(object):
""" """
Variable = graph.Variable #the type that will be created by call to make_variable. Variable = graph.Variable # the type that will be created by call to make_variable.
Constant = graph.Constant #the type that will be created by call to make_constant Constant = graph.Constant # the type that will be created by call to make_constant
def filter(self, data, strict=False, allow_downcast=None): def filter(self, data, strict=False, allow_downcast=None):
"""Required: Return data or an appropriately wrapped/converted data. """Required: Return data or an appropriately wrapped/converted data.
......
theano/gof/unify.py
浏览文件 @ 3089deda
...@@ -71,7 +71,7 @@ class NotVariable(Variable): ...@@ -71,7 +71,7 @@ class NotVariable(Variable):
self.name = name self.name = name
self.not_options = not_options self.not_options = not_options
class VariableInList: # not a subclass of Variable class VariableInList: # not a subclass of Variable
""" """
This special kind of variable is matched against a list and unifies This special kind of variable is matched against a list and unifies
an inner Variable to an OrVariable of the values in the list. For an inner Variable to an OrVariable of the values in the list. For
...@@ -341,9 +341,9 @@ def unify_walk(v, o, U): ...@@ -341,9 +341,9 @@ def unify_walk(v, o, U):
""" """
best_v = U[v] best_v = U[v]
if v is not best_v: if v is not best_v:
return unify_walk(o, best_v, U) # reverse argument order so if o is a Variable this block of code is run again return unify_walk(o, best_v, U) # reverse argument order so if o is a Variable this block of code is run again
else: else:
return FALL_THROUGH # call the next version of unify_walk that matches the type signature return FALL_THROUGH # call the next version of unify_walk that matches the type signature
################################ ################################
...@@ -415,9 +415,9 @@ def unify_merge(v, o, U): ...@@ -415,9 +415,9 @@ def unify_merge(v, o, U):
""" """
best_v = U[v] best_v = U[v]
if v is not best_v: if v is not best_v:
return unify_merge(o, best_v, U) # reverse argument order so if o is a Variable this block of code is run again return unify_merge(o, best_v, U) # reverse argument order so if o is a Variable this block of code is run again
else: else:
return FALL_THROUGH # call the next version of unify_walk that matches the type signature return FALL_THROUGH # call the next version of unify_walk that matches the type signature
################################ ################################
......
theano/misc/check_duplicate_key.py
浏览文件 @ 3089deda
...@@ -13,7 +13,7 @@ if len(sys.argv)>1: ...@@ -13,7 +13,7 @@ if len(sys.argv)>1:
else: else:
dirs = os.listdir(theano.config.compiledir) dirs = os.listdir(theano.config.compiledir)
dirs = [os.path.join(theano.config.compiledir, d) for d in dirs] dirs = [os.path.join(theano.config.compiledir, d) for d in dirs]
keys = {} # key -> nb seen keys = {} # key -> nb seen
mods = {} mods = {}
for dir in dirs: for dir in dirs:
...@@ -49,13 +49,13 @@ if DISPLAY_DUPLICATE_KEYS: ...@@ -49,13 +49,13 @@ if DISPLAY_DUPLICATE_KEYS:
if v > 1: if v > 1:
print "Duplicate key (%i copies): %s" % (v, cPickle.loads(k)) print "Duplicate key (%i copies): %s" % (v, cPickle.loads(k))
nbs_keys = {} # nb seen -> now many key nbs_keys = {} # nb seen -> now many key
for val in keys.values(): for val in keys.values():
nbs_keys.setdefault(val, 0) nbs_keys.setdefault(val, 0)
nbs_keys[val]+=1 nbs_keys[val]+=1
nbs_mod = {} # nb seen -> how many key nbs_mod = {} # nb seen -> how many key
nbs_mod_to_key = {} #nb seen -> keys nbs_mod_to_key = {} # nb seen -> keys
more_than_one = 0 more_than_one = 0
for mod, kk in mods.iteritems(): for mod, kk in mods.iteritems():
val = len(kk) val = len(kk)
......
theano/misc/cudamat_utils.py
浏览文件 @ 3089deda
...@@ -71,7 +71,7 @@ try: ...@@ -71,7 +71,7 @@ try:
cm_mat.on_host = 0 cm_mat.on_host = 0
cm_mat.on_device = 1 cm_mat.on_device = 1
cm_mat.is_trans = 0 cm_mat.is_trans = 0
cm_mat.owns_data = 0 # <-- note: cm_mat dosen't owe the data; x does. So x will delete it. cm_mat.owns_data = 0 # <-- note: cm_mat dosen't owe the data; x does. So x will delete it.
# x.gpudata is a long. We need a pointer to a float. cast. # x.gpudata is a long. We need a pointer to a float. cast.
...@@ -80,12 +80,12 @@ try: ...@@ -80,12 +80,12 @@ try:
px = cudamat.CUDAMatrix(cm_mat) px = cudamat.CUDAMatrix(cm_mat)
px._base = x # x won't be __del__'ed as long as px is around. px._base = x # x won't be __del__'ed as long as px is around.
px.mat_on_host = False # let cudamat know that we don't have a numpy px.mat_on_host = False # let cudamat know that we don't have a numpy
# array attached. # array attached.
return px return px
......
theano/misc/gnumpy_utils.py
浏览文件 @ 3089deda
...@@ -63,7 +63,7 @@ try: ...@@ -63,7 +63,7 @@ try:
cm_mat.on_host = 0 cm_mat.on_host = 0
cm_mat.on_device = 1 cm_mat.on_device = 1
cm_mat.is_trans = 0 cm_mat.is_trans = 0
cm_mat.owns_data = 0 # <-- note: cm_mat dosen't owe the data; x does. So x will delete it. cm_mat.owns_data = 0 # <-- note: cm_mat dosen't owe the data; x does. So x will delete it.
# x.gpudata is a long. We need a pointer to a float. cast. # x.gpudata is a long. We need a pointer to a float. cast.
...@@ -72,12 +72,12 @@ try: ...@@ -72,12 +72,12 @@ try:
px = cudamat.CUDAMatrix(cm_mat) px = cudamat.CUDAMatrix(cm_mat)
px._base = x # x won't be freed if the cudamat object isn't freed. px._base = x # x won't be freed if the cudamat object isn't freed.
px.mat_on_host = False # let cudamat know that we don't have a numpy px.mat_on_host = False # let cudamat know that we don't have a numpy
# array attached. # array attached.
# Note how gnumpy tracks its cudamat objects: it moves things to the # Note how gnumpy tracks its cudamat objects: it moves things to the
......
theano/misc/hooks/check_whitespace.py
浏览文件 @ 3089deda
...@@ -112,7 +112,7 @@ def run_mercurial_command(hg_command): ...@@ -112,7 +112,7 @@ def run_mercurial_command(hg_command):
def parse_stdout_filelist(hg_out_filelist): def parse_stdout_filelist(hg_out_filelist):
files = hg_out_filelist.split() files = hg_out_filelist.split()
files = [f.strip(string.whitespace + "'") for f in files] files = [f.strip(string.whitespace + "'") for f in files]
files = filter(operator.truth, files) # get rid of empty entries files = filter(operator.truth, files) # get rid of empty entries
return files return files
def changed_files(): def changed_files():
...@@ -221,14 +221,14 @@ def main(argv=None): ...@@ -221,14 +221,14 @@ def main(argv=None):
else: else:
# parsing succeeded, it is safe to check indentation # parsing succeeded, it is safe to check indentation
if not args.no_indentation: if not args.no_indentation:
was_clean = None # unknown was_clean = None # unknown
# only calculate was_clean if it will matter to us # only calculate was_clean if it will matter to us
if args.incremental or args.incremental_with_patch: if args.incremental or args.incremental_with_patch:
if filename in changed_filenames: if filename in changed_filenames:
old_file_contents = get_file_contents(filename, revision=parent_commit()) old_file_contents = get_file_contents(filename, revision=parent_commit())
was_clean = get_correct_indentation_diff(old_file_contents, "") is None was_clean = get_correct_indentation_diff(old_file_contents, "") is None
else: else:
was_clean = True # by default -- it was newly added and thus had no prior problems was_clean = True # by default -- it was newly added and thus had no prior problems
check_indentation = was_clean or not args.incremental check_indentation = was_clean or not args.incremental
if check_indentation: if check_indentation:
......
theano/misc/pycuda_utils.py
浏览文件 @ 3089deda
...@@ -60,6 +60,6 @@ def to_cudandarray(x): ...@@ -60,6 +60,6 @@ def to_cudandarray(x):
for i in x.shape[::-1][:-1]: for i in x.shape[::-1][:-1]:
strides.append(strides[-1]*i) strides.append(strides[-1]*i)
strides = tuple(strides[::-1]) strides = tuple(strides[::-1])
ptr = int(x.gpudata) # in pycuda trunk, y.ptr also works, which is a little cleaner ptr = int(x.gpudata) # in pycuda trunk, y.ptr also works, which is a little cleaner
z = cuda.from_gpu_pointer(ptr, x.shape, strides, x) z = cuda.from_gpu_pointer(ptr, x.shape, strides, x)
return z return z
theano/sandbox/cuda/basic_ops.py
浏览文件 @ 3089deda
...@@ -12,7 +12,7 @@ from theano.compat.six import StringIO ...@@ -12,7 +12,7 @@ from theano.compat.six import StringIO
from theano.gradient import grad_undefined from theano.gradient import grad_undefined
from theano.scalar import Scalar from theano.scalar import Scalar
scal = scalar # somewhere scalar gets reassigned to be a function scal = scalar # somewhere scalar gets reassigned to be a function
from theano.compat.python2x import all, any from theano.compat.python2x import all, any
...@@ -3415,7 +3415,7 @@ class GpuAlloc(GpuAllocEmpty): ...@@ -3415,7 +3415,7 @@ class GpuAlloc(GpuAllocEmpty):
# If the output is a constant, it will have to be deepcopied # If the output is a constant, it will have to be deepcopied
# each time the function is called. So we do not fold. # each time the function is called. So we do not fold.
return False return False
elif (#The following ops work inplace of their input id 0. elif ( # The following ops work inplace of their input id 0.
client[1] == 0 and client[1] == 0 and
isinstance(client[0].op, ( isinstance(client[0].op, (
# Ops that will work inplace on the Alloc. So if they # Ops that will work inplace on the Alloc. So if they
......
theano/sandbox/cuda/elemwise.py
浏览文件 @ 3089deda
...@@ -260,7 +260,7 @@ class NaiveAlgo(object): ...@@ -260,7 +260,7 @@ class NaiveAlgo(object):
, nodename + '_scalar_' , nodename + '_scalar_'
, get_str_list_logical_scalar(node, value_str='value0[%i]') , get_str_list_logical_scalar(node, value_str='value0[%i]')
, ['ii_o%i_data[0]'%ipos for ipos, i in enumerate(node.outputs)] , ['ii_o%i_data[0]'%ipos for ipos, i in enumerate(node.outputs)]
, sub=dict(fail='return;')) #TODO: set a failure code somehow!!! , sub=dict(fail='return;')) # TODO: set a failure code somehow!!!
print >> sio, " ", task_code print >> sio, " ", task_code
print >> sio, " }" * nd print >> sio, " }" * nd
...@@ -398,7 +398,7 @@ class NaiveAlgo(object): ...@@ -398,7 +398,7 @@ class NaiveAlgo(object):
, nodename + '_scalar_' , nodename + '_scalar_'
, ['i%i_data_%i[0]'%(ipos, d) for ipos, i in enumerate(node.inputs)] , ['i%i_data_%i[0]'%(ipos, d) for ipos, i in enumerate(node.inputs)]
, ['o%i_data_%i[0]'%(ipos, d) for ipos, i in enumerate(node.outputs)] , ['o%i_data_%i[0]'%(ipos, d) for ipos, i in enumerate(node.outputs)]
, sub=dict(fail='return;')) #TODO: set a failure code somehow!!! , sub=dict(fail='return;')) # TODO: set a failure code somehow!!!
if nd == 4: if nd == 4:
decl_shared_stride(n_in, n_out, nd) decl_shared_stride(n_in, n_out, nd)
...@@ -473,7 +473,7 @@ class NaiveAlgo(object): ...@@ -473,7 +473,7 @@ class NaiveAlgo(object):
#, ['i%i_data[i]'%ipos for ipos, i in enumerate(node.inputs)] #, ['i%i_data[i]'%ipos for ipos, i in enumerate(node.inputs)]
, get_str_list_logical_scalar(node, data_str='i%i_data[i]') , get_str_list_logical_scalar(node, data_str='i%i_data[i]')
, ['o%i_data[i]'%ipos for ipos, i in enumerate(node.outputs)] , ['o%i_data[i]'%ipos for ipos, i in enumerate(node.outputs)]
, sub=dict(fail='return;')) #TODO: set a failure code somehow!!! , sub=dict(fail='return;')) # TODO: set a failure code somehow!!!
print >> sio, " ", task_code print >> sio, " ", task_code
print >> sio, " }" print >> sio, " }"
print >> sio, "}" print >> sio, "}"
...@@ -844,9 +844,9 @@ nd_collapse_[i]=0; ...@@ -844,9 +844,9 @@ nd_collapse_[i]=0;
launch_General(nodename, scalar_op, i, self.sync) launch_General(nodename, scalar_op, i, self.sync)
print >> sio, " } break;" print >> sio, " } break;"
print >> sio, "}"#end case print >> sio, "}" # end case
print >> sio, "return -2;" # should not get to this point print >> sio, "return -2;" # should not get to this point
print >> sio, "}"#end fct print >> sio, "}" # end fct
# N.B. cudaGetLastError is called by c_code # N.B. cudaGetLastError is called by c_code
return sio.getvalue() return sio.getvalue()
......
theano/sandbox/cuda/tests/test_basic_ops.py
浏览文件 @ 3089deda
...@@ -87,7 +87,7 @@ def test_careduce(): ...@@ -87,7 +87,7 @@ def test_careduce():
((1, 2), (1,)), ((1, 2), (1,)),
((100, 3, 1300), [1]), ((100, 3, 1300), [1]),
((0,), [0]), ((5,), [0]), ((0,), [0]), ((5,), [0]),
((0, 0), [0, 1]), ((1, 0), [0, 1]), ((5, 4), [0, 1]), ((33, 31), [0, 1]), ((5, 4), [1]), ((5, 4), [0]),#need something bigger then 32 for some opt test. ((0, 0), [0, 1]), ((1, 0), [0, 1]), ((5, 4), [0, 1]), ((33, 31), [0, 1]), ((5, 4), [1]), ((5, 4), [0]), # need something bigger then 32 for some opt test.
((5, 4, 3), [0]), ((5, 4, 3), [1]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [1, 2]), ((5, 4, 3), [0, 1, 2]), ((5, 4, 3), [0]), ((5, 4, 3), [1]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [1, 2]), ((5, 4, 3), [0, 1, 2]),
((0, 0, 0, 0), [0, 1, 2, 3]), ((0, 0, 0, 0), [0, 1, 2, 3]),
((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]), ((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]),
...@@ -95,57 +95,57 @@ def test_careduce(): ...@@ -95,57 +95,57 @@ def test_careduce():
((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]), ((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]),
# test shape bigger then 4096 on each dimension to make sure that we work correctly when we don't have enough thread/block in each dimensions # test shape bigger then 4096 on each dimension to make sure that we work correctly when we don't have enough thread/block in each dimensions
((4100, 3), [0]), ((3, 4101), [0]),#10 ((4100, 3), [0]), ((3, 4101), [0]), # 10
((1024, 33), [0]), ((33, 1024), [0]),#10 ((1024, 33), [0]), ((33, 1024), [0]), # 10
((1025, 33), [0]), ((33, 1025), [0]),#10 ((1025, 33), [0]), ((33, 1025), [0]), # 10
((4100, 3), [1]), ((3, 4101), [1]),#01 ((4100, 3), [1]), ((3, 4101), [1]), # 01
((1024, 33), [1]), ((33, 1024), [1]),#01 ((1024, 33), [1]), ((33, 1024), [1]), # 01
((1025, 33), [1]), ((33, 1025), [1]),#01 ((1025, 33), [1]), ((33, 1025), [1]), # 01
((4100, 3), [0, 1]), ((3, 4101), [0, 1]),#11 ((4100, 3), [0, 1]), ((3, 4101), [0, 1]), # 11
((1024, 33), [0, 1]), ((33, 1024), [0, 1]),#01 ((1024, 33), [0, 1]), ((33, 1024), [0, 1]), # 01
((1025, 33), [0, 1]), ((33, 1025), [0, 1]),#01 ((1025, 33), [0, 1]), ((33, 1025), [0, 1]), # 01
((4100, 4, 3), [0]), ((5, 4100, 3), [0]), ((5, 4, 4100), [0]), ((3, 65536, 1), [0]),#100 ((4100, 4, 3), [0]), ((5, 4100, 3), [0]), ((5, 4, 4100), [0]), ((3, 65536, 1), [0]), # 100
((4100, 4, 3), [1]), ((5, 4100, 3), [1]), ((5, 4, 4100), [1]),#010 ((4100, 4, 3), [1]), ((5, 4100, 3), [1]), ((5, 4, 4100), [1]), # 010
((4100, 4, 3), [2]), ((5, 4100, 3), [2]), ((5, 4, 4100), [2]),#001 ((4100, 4, 3), [2]), ((5, 4100, 3), [2]), ((5, 4, 4100), [2]), # 001
((4100, 4, 3), [0, 1]), ((5, 4100, 3), [0, 1]), ((5, 4, 4100), [0, 1]),#110 ((4100, 4, 3), [0, 1]), ((5, 4100, 3), [0, 1]), ((5, 4, 4100), [0, 1]), # 110
((4100, 4, 3), [1, 2]), ((5, 4100, 3), [1, 2]), ((5, 4, 4100), [1, 2]),#011 ((4100, 4, 3), [1, 2]), ((5, 4100, 3), [1, 2]), ((5, 4, 4100), [1, 2]), # 011
#((4100,4,3),[0,2]),((5,4100,3),[0,2]),((5,4,4100),[0,2]),#101 ##not implemented #((4100,4,3),[0,2]),((5,4100,3),[0,2]),((5,4,4100),[0,2]),#101 ##not implemented
((4100, 4, 3), [0, 1, 2]), ((5, 4100, 3), [0, 1, 2]), ((5, 4, 4100), [0, 1, 2]),#111 ((4100, 4, 3), [0, 1, 2]), ((5, 4100, 3), [0, 1, 2]), ((5, 4, 4100), [0, 1, 2]), # 111
((65, 4, 3), [0, 1, 2]), ((5, 65, 3), [0, 1, 2]), ((5, 4, 65), [0, 1, 2]),#111 ((65, 4, 3), [0, 1, 2]), ((5, 65, 3), [0, 1, 2]), ((5, 4, 65), [0, 1, 2]), # 111
((4100, 4, 3, 2), [2, 3]), ((4, 4100, 3, 2), [2, 3]), ((4, 3, 4100, 2), [2, 3]), ((4, 3, 2, 4100), [2, 3]),#0011 ((4100, 4, 3, 2), [2, 3]), ((4, 4100, 3, 2), [2, 3]), ((4, 3, 4100, 2), [2, 3]), ((4, 3, 2, 4100), [2, 3]), # 0011
((4100, 4, 3, 2), [1, 3]), ((4, 4100, 3, 2), [1, 3]), ((4, 3, 4100, 2), [1, 3]), ((4, 3, 2, 4100), [1, 3]),#0101 ((4100, 4, 3, 2), [1, 3]), ((4, 4100, 3, 2), [1, 3]), ((4, 3, 4100, 2), [1, 3]), ((4, 3, 2, 4100), [1, 3]), # 0101
((4100, 4, 3, 2), [0, 2, 3]), ((4, 4100, 3, 2), [0, 2, 3]), ((4, 3, 4100, 2), [0, 2, 3]),#((4,3,2,4100),[0,2,3]),#1011 ((4100, 4, 3, 2), [0, 2, 3]), ((4, 4100, 3, 2), [0, 2, 3]), ((4, 3, 4100, 2), [0, 2, 3]), # ((4,3,2,4100),[0,2,3]),#1011
((4100, 4, 3, 2), [1, 2, 3]), ((4, 4100, 3, 2), [1, 2, 3]), ((4, 3, 4100, 2), [1, 2, 3]), ((4, 3, 2, 4100), [1, 2, 3]),#0111 ((4100, 4, 3, 2), [1, 2, 3]), ((4, 4100, 3, 2), [1, 2, 3]), ((4, 3, 4100, 2), [1, 2, 3]), ((4, 3, 2, 4100), [1, 2, 3]), # 0111
((65, 4, 3, 2), [1, 2, 3]), ((4, 65, 3, 2), [1, 2, 3]), ((4, 3, 65, 2), [1, 2, 3]), ((4, 3, 2, 65), [1, 2, 3]),#0111 ((65, 4, 3, 2), [1, 2, 3]), ((4, 65, 3, 2), [1, 2, 3]), ((4, 3, 65, 2), [1, 2, 3]), ((4, 3, 2, 65), [1, 2, 3]), # 0111
((4100, 2, 3, 4), [0, 1, 2, 3]), ((2, 4100, 3, 4), [0, 1, 2, 3]), ((2, 3, 4100, 4), [0, 1, 2, 3]), ((2, 3, 4, 4100), [0, 1, 2, 3]), ((128, 1, 3, 3), [0, 1, 2, 3]),#1111 ((4100, 2, 3, 4), [0, 1, 2, 3]), ((2, 4100, 3, 4), [0, 1, 2, 3]), ((2, 3, 4100, 4), [0, 1, 2, 3]), ((2, 3, 4, 4100), [0, 1, 2, 3]), ((128, 1, 3, 3), [0, 1, 2, 3]), # 1111
# test pattern implemented by reshape # test pattern implemented by reshape
((4100, 4, 3, 2), [0]), ((4, 4100, 3, 2), [0]), ((4, 3, 4100, 2), [0]), ((4, 3, 2, 4100), [0]),#1000 ((4100, 4, 3, 2), [0]), ((4, 4100, 3, 2), [0]), ((4, 3, 4100, 2), [0]), ((4, 3, 2, 4100), [0]), # 1000
((4100, 4, 3, 2), [1]), ((4, 4100, 3, 2), [1]), ((4, 3, 4100, 2), [1]), ((4, 3, 2, 4100), [1]),#0100 ((4100, 4, 3, 2), [1]), ((4, 4100, 3, 2), [1]), ((4, 3, 4100, 2), [1]), ((4, 3, 2, 4100), [1]), # 0100
((4100, 4, 3, 2), [2]), ((4, 4100, 3, 2), [2]), ((4, 3, 4100, 2), [2]), ((4, 3, 2, 4100), [2]),#0010 ((4100, 4, 3, 2), [2]), ((4, 4100, 3, 2), [2]), ((4, 3, 4100, 2), [2]), ((4, 3, 2, 4100), [2]), # 0010
((4100, 4, 3, 2), [3]), ((4, 4100, 3, 2), [3]), ((4, 3, 4100, 2), [3]), ((4, 3, 2, 4100), [3]),#0001 ((4100, 4, 3, 2), [3]), ((4, 4100, 3, 2), [3]), ((4, 3, 4100, 2), [3]), ((4, 3, 2, 4100), [3]), # 0001
# reduce over 2d # reduce over 2d
((4100, 4, 3, 2), [1, 2]), ((4, 4100, 3, 2), [1, 2]), ((4, 3, 4100, 2), [1, 2]), ((4, 3, 2, 4100), [1, 2]),#0110 ((4100, 4, 3, 2), [1, 2]), ((4, 4100, 3, 2), [1, 2]), ((4, 3, 4100, 2), [1, 2]), ((4, 3, 2, 4100), [1, 2]), # 0110
# ((4100,4,3,2),[0,3]),((4,4100,3,2),[0,3]),((4,3,4100,2),[0,3]),((4,3,2,4100),[0,3]),#1001 need 101 # ((4100,4,3,2),[0,3]),((4,4100,3,2),[0,3]),((4,3,4100,2),[0,3]),((4,3,2,4100),[0,3]),#1001 need 101
# ((4100,4,3,2),[0,2]),((4,4100,3,2),[0,2]),((4,3,4100,2),[0,2]),((4,3,2,4100),[0,2]),#1010 not implemented # ((4100,4,3,2),[0,2]),((4,4100,3,2),[0,2]),((4,3,4100,2),[0,2]),((4,3,2,4100),[0,2]),#1010 not implemented
((4100, 4, 3, 2), [0, 1]), ((4, 4100, 3, 2), [0, 1]), ((4, 3, 4100, 2), [0, 1]), ((4, 3, 2, 4100), [0, 1]),#1100 ((4100, 4, 3, 2), [0, 1]), ((4, 4100, 3, 2), [0, 1]), ((4, 3, 4100, 2), [0, 1]), ((4, 3, 2, 4100), [0, 1]), # 1100
# reduce over 3d # reduce over 3d
# 3d not tested: 1101, 1110, 1111 # 3d not tested: 1101, 1110, 1111
# ((4100,4,3,2),[0,1,3]),((4,4100,3,2),[0,1,3]),((4,3,4100,2),[0,1,3]),((4,3,2,4100),[0,1,3]),#1101 need 101 # ((4100,4,3,2),[0,1,3]),((4,4100,3,2),[0,1,3]),((4,3,4100,2),[0,1,3]),((4,3,2,4100),[0,1,3]),#1101 need 101
((4100, 4, 3, 2), [0, 1, 2]), ((4, 4100, 3, 2), [0, 1, 2]), ((4, 3, 4100, 2), [0, 1, 2]), ((4, 3, 2, 4100), [0, 1, 2]),#1110 ((4100, 4, 3, 2), [0, 1, 2]), ((4, 4100, 3, 2), [0, 1, 2]), ((4, 3, 4100, 2), [0, 1, 2]), ((4, 3, 2, 4100), [0, 1, 2]), # 1110
# reduce over 4d # reduce over 4d
((4100, 4, 3, 2), [0]), ((4, 4100, 3, 2), [0]), ((4, 3, 4100, 2), [0]), ((4, 3, 2, 4100), [0]),#1111 ((4100, 4, 3, 2), [0]), ((4, 4100, 3, 2), [0]), ((4, 3, 4100, 2), [0]), ((4, 3, 2, 4100), [0]), # 1111
# reduce over 5d # reduce over 5d
((1100, 2, 3, 4, 5), [0, 1, 2, 3, 4]), ((2, 1100, 3, 4, 5), [0, 1, 2, 3, 4]), ((2, 3, 1100, 4, 5), [0, 1, 2, 3, 4]), ((2, 3, 4, 1100, 5), [0, 1, 2, 3, 4]), ((2, 3, 4, 5, 1100), [0, 1, 2, 3, 4]),#11111 ((1100, 2, 3, 4, 5), [0, 1, 2, 3, 4]), ((2, 1100, 3, 4, 5), [0, 1, 2, 3, 4]), ((2, 3, 1100, 4, 5), [0, 1, 2, 3, 4]), ((2, 3, 4, 1100, 5), [0, 1, 2, 3, 4]), ((2, 3, 4, 5, 1100), [0, 1, 2, 3, 4]), # 11111
]: ]:
...@@ -221,7 +221,7 @@ def test_careduce(): ...@@ -221,7 +221,7 @@ def test_careduce():
# test with dimshuffle # test with dimshuffle
# we shuffle the 2 outer dims. # we shuffle the 2 outer dims.
for shape, pattern in [#((5,),[0]), for shape, pattern in [ # ((5,),[0]),
((5, 4), [0, 1]), ((5, 4), [0]), ((5, 4), [0, 1]), ((5, 4), [0]),
((5, 4, 3), [0]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [0, 1, 2]), ((5, 4, 3), [0]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [0, 1, 2]),
((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]),
...@@ -1260,7 +1260,7 @@ test_shared_options = theano.tensor.tests.test_sharedvar.makeSharedTester( ...@@ -1260,7 +1260,7 @@ test_shared_options = theano.tensor.tests.test_sharedvar.makeSharedTester(
shared_constructor_=tcn.shared_constructor, shared_constructor_=tcn.shared_constructor,
dtype_='float32', dtype_='float32',
get_value_borrow_true_alias_=True, get_value_borrow_true_alias_=True,
shared_borrow_true_alias_=True,#True when the original value is already a CudaNdarray! shared_borrow_true_alias_=True, # True when the original value is already a CudaNdarray!
set_value_borrow_true_alias_=True, set_value_borrow_true_alias_=True,
set_value_inplace_=True, set_value_inplace_=True,
set_cast_value_inplace_=False, set_cast_value_inplace_=False,
......
theano/sandbox/cuda/tests/test_bench_loopfusion.py
浏览文件 @ 3089deda
...@@ -93,7 +93,7 @@ class Kouh2008(object): ...@@ -93,7 +93,7 @@ class Kouh2008(object):
p_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='p') p_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='p')
q_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='q') q_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='q')
r_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='r') r_unbounded = shared_uniform(low=-0.1, high=0.1, size=(n_out,), name='r')
k_unbounded = shared_uniform(low=-0.2, high=0.2, size=(n_out,), name='k') # biases k_unbounded = shared_uniform(low=-0.2, high=0.2, size=(n_out,), name='k') # biases
p = tensor.nnet.sigmoid(p_unbounded) * e_range_mag + e_range_low p = tensor.nnet.sigmoid(p_unbounded) * e_range_mag + e_range_low
q = tensor.nnet.sigmoid(q_unbounded) * e_range_mag + e_range_low q = tensor.nnet.sigmoid(q_unbounded) * e_range_mag + e_range_low
...@@ -171,7 +171,7 @@ class Kouh2008(object): ...@@ -171,7 +171,7 @@ class Kouh2008(object):
rval = cls.new_expbounds(rng, x_list, n_out, dtype=dtype, params=f_list + b_list, rval = cls.new_expbounds(rng, x_list, n_out, dtype=dtype, params=f_list + b_list,
exponent_range=exponent_range) exponent_range=exponent_range)
rval.f_list = f_list rval.f_list = f_list
rval.input = input #add the input to the returned object rval.input = input # add the input to the returned object
rval.filter_l1 = sum(abs(fi).sum() for fi in f_list) rval.filter_l1 = sum(abs(fi).sum() for fi in f_list)
rval.filter_l2_sqr = sum((fi**2).sum() for fi in f_list) rval.filter_l2_sqr = sum((fi**2).sum() for fi in f_list)
return rval return rval
...@@ -210,18 +210,18 @@ class Kouh2008(object): ...@@ -210,18 +210,18 @@ class Kouh2008(object):
out_c_high = out_c_low + filter_shape[1] out_c_high = out_c_low + filter_shape[1]
out_tile = out_array[out_r_low:out_r_high, out_c_low:out_c_high, :] out_tile = out_array[out_r_low:out_r_high, out_c_low:out_c_high, :]
if c % 3 == 0: # linear filter if c % 3 == 0: # linear filter
if w_col < w.shape[1]: if w_col < w.shape[1]:
out_tile[...] = pixel_range(w[:, w_col]).reshape(filter_shape+(1,)) out_tile[...] = pixel_range(w[:, w_col]).reshape(filter_shape+(1,))
w_col += 1 w_col += 1
if c % 3 == 1: # E filters if c % 3 == 1: # E filters
if w_col < w.shape[1]: if w_col < w.shape[1]:
# filters after the 3rd do not get rendered, but are skipped over. # filters after the 3rd do not get rendered, but are skipped over.
# there are only 3 colour channels. # there are only 3 colour channels.
for i in xrange(min(self.n_E_quadratic, 3)): for i in xrange(min(self.n_E_quadratic, 3)):
out_tile[:, :, i] = pixel_range(w[:, w_col+i]).reshape(filter_shape) out_tile[:, :, i] = pixel_range(w[:, w_col+i]).reshape(filter_shape)
w_col += self.n_E_quadratic w_col += self.n_E_quadratic
if c % 3 == 2: # S filters if c % 3 == 2: # S filters
if w_col < w.shape[1]: if w_col < w.shape[1]:
# filters after the 3rd do not get rendered, but are skipped over. # filters after the 3rd do not get rendered, but are skipped over.
# there are only 3 colour channels. # there are only 3 colour channels.
...@@ -244,7 +244,7 @@ class Config(object): ...@@ -244,7 +244,7 @@ class Config(object):
n_terms = 4 n_terms = 4
ft_lr_t0 = 3e-3 ft_lr_t0 = 3e-3
ft_t_decay = 0 # 50 * 5000 # (units of minibatches) by this N'th pass through the training set ft_t_decay = 0 # 50 * 5000 # (units of minibatches) by this N'th pass through the training set
ft_lr_t_decay = 1e-3 # we will have this learning rate ft_lr_t_decay = 1e-3 # we will have this learning rate
ft_cost_classif_l1 = 0 ft_cost_classif_l1 = 0
ft_cost_classif_l2 = 0 ft_cost_classif_l2 = 0
...@@ -255,11 +255,11 @@ class Config(object): ...@@ -255,11 +255,11 @@ class Config(object):
ft_cost_in_l1_w = 0 ft_cost_in_l1_w = 0
ft_cost_in_l2_w = 0 ft_cost_in_l2_w = 0
ft_limit_iters = -1 ft_limit_iters = -1
ft_limit_walltime = 0 # in seconds 60*60*1 #1 hour ft_limit_walltime = 0 # in seconds 60*60*1 #1 hour
ft_batchsize = 30 ft_batchsize = 30
ft_epoch_len = 50000 ft_epoch_len = 50000
ft_status_interval = 50 #property( lambda s:s.ft_epoch_len/s.ft_batchsize) ft_status_interval = 50 # property( lambda s:s.ft_epoch_len/s.ft_batchsize)
ft_validation_interval = property( lambda s: s.ft_epoch_len/s.ft_batchsize) ft_validation_interval = property( lambda s: s.ft_epoch_len/s.ft_batchsize)
ft_ntrain_limit = 0 ft_ntrain_limit = 0
ft_test_lag1 = True ft_test_lag1 = True
......
theano/sandbox/cuda/tests/test_conv_cuda_ndarray.py
浏览文件 @ 3089deda
...@@ -337,28 +337,28 @@ def get_valid_shapes(): ...@@ -337,28 +337,28 @@ def get_valid_shapes():
, ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1))
, ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize , ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize
, ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image , ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image
, ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image, non-square kern , ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image, non-square kern
, ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim , ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim
, ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one , ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one
, ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1 , ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1
, ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights , ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights
, ((60, 20, 28, 28), (10, 20, 5, 5), (1, 1), (2, 2), (1, 1))#added a test case that fail from test_nnet.py.test_conv_nnet2 , ((60, 20, 28, 28), (10, 20, 5, 5), (1, 1), (2, 2), (1, 1)) # added a test case that fail from test_nnet.py.test_conv_nnet2
, ((10, 5, 28, 28), (10, 5, 5, 5), (1, 1), (2, 2), (1, 1))#test precedent but reduced that triger the error , ((10, 5, 28, 28), (10, 5, 5, 5), (1, 1), (2, 2), (1, 1)) # test precedent but reduced that triger the error
# Test more than maxThreadsDim0 # Test more than maxThreadsDim0
, ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
, ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
] ]
shapes += [ ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers shapes += [ ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 1 layers
, ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers , ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 2 layers
, ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 1 full , ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 1 full
, ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid , ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 2 valid
# , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
, ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers , ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 1 layers
, ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers , ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 2 layers
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 full , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 full
# , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1 # , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1
# , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2 # , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2
] ]
...@@ -461,36 +461,36 @@ def _test_full(cls, mode=None, version=[-1], extra_shapes=[]): ...@@ -461,36 +461,36 @@ def _test_full(cls, mode=None, version=[-1], extra_shapes=[]):
, ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1))
, ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize , ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize
, ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image , ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image
, ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image, non-square kern , ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image, non-square kern
, ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim , ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim
, ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one , ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one
, ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1 , ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1
, ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights , ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights
# other test # other test
, ((3, 1, 1, 1), (2, 1, 5, 3), (1, 1), (1, 1), (1, 1))#kernel bigger then image , ((3, 1, 1, 1), (2, 1, 5, 3), (1, 1), (1, 1), (1, 1)) # kernel bigger then image
, ((3, 2, 1, 1), (4, 2, 1, 1), (1, 1), (1, 1), (1, 1)) , ((3, 2, 1, 1), (4, 2, 1, 1), (1, 1), (1, 1), (1, 1))
, ((3, 2, 4, 4), (4, 2, 2, 6), (1, 1), (1, 1), (1, 1)) , ((3, 2, 4, 4), (4, 2, 2, 6), (1, 1), (1, 1), (1, 1))
, ((3, 2, 4, 4), (4, 2, 8, 6), (1, 1), (1, 1), (1, 1))#kernel bigger then image , ((3, 2, 4, 4), (4, 2, 8, 6), (1, 1), (1, 1), (1, 1)) # kernel bigger then image
, ((4, 2, 10, 10), (3, 2, 2, 12), (1, 1), (1, 1), (1, 1)) , ((4, 2, 10, 10), (3, 2, 2, 12), (1, 1), (1, 1), (1, 1))
] ]
shapes += [ shapes += [
# ((60,1,28,28),(20,1,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers # ((60,1,28,28),(20,1,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers
# , ((60,20,12,12),(30,20,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers # , ((60,20,12,12),(30,20,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers
((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 1 full ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 1 full
# , ((20,60,12,12),(30,60,8,8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((20,60,12,12),(30,60,8,8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
# , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
# , ((10,1,64,64),(20,1,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers # , ((10,1,64,64),(20,1,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers
# , ((10,20,29,29),(30,20,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers # , ((10,20,29,29),(30,20,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 full , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 full
# , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1 # , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1
# , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2 # , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2
# Test more than maxThreadsDim0 # Test more than maxThreadsDim0
, ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
, ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
, ((1, 1, 44800, 1), (6, 1, 1, 1), (1, 1), (1, 1), (1, 1))#This caused crash , ((1, 1, 44800, 1), (6, 1, 1, 1), (1, 1), (1, 1), (1, 1)) # This caused crash
] ]
verbose = 0 verbose = 0
...@@ -869,43 +869,43 @@ def benchmark(): ...@@ -869,43 +869,43 @@ def benchmark():
shapes_valid = [ shapes_valid = [
# test_lenet_28 shape # test_lenet_28 shape
((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1))#valid ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 60, 28, 28), (20, 60, 24, 24), (1, 1), (1, 1), (1, 1))#valid , ((1, 60, 28, 28), (20, 60, 24, 24), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_32 shape # test_lenet_32 shape
, ((20, 60, 14, 14), (30, 60, 10, 10), (1, 1), (1, 1), (1, 1))#valid , ((20, 60, 14, 14), (30, 60, 10, 10), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 20, 14, 14), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 20, 14, 14), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 1, 32, 32), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 1, 32, 32), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 60, 32, 32), (20, 60, 28, 28), (1, 1), (1, 1), (1, 1))#valid , ((1, 60, 32, 32), (20, 60, 28, 28), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_64 shape # test_lenet_64 shape
, ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 10, 29, 29), (30, 10, 23, 23), (1, 1), (1, 1), (1, 1))#valid , ((20, 10, 29, 29), (30, 10, 23, 23), (1, 1), (1, 1), (1, 1)) # valid
, ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 10, 64, 64), (20, 10, 58, 58), (1, 1), (1, 1), (1, 1))#valid , ((1, 10, 64, 64), (20, 10, 58, 58), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_108 shape # test_lenet_108 shape
, ((10, 20, 51, 51), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 20, 51, 51), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 10, 51, 51), (30, 10, 45, 45), (1, 1), (1, 1), (1, 1))#valid , ((20, 10, 51, 51), (30, 10, 45, 45), (1, 1), (1, 1), (1, 1)) # valid
, ((10, 1, 108, 108), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 1, 108, 108), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 10, 108, 108), (20, 10, 102, 102), (1, 1), (1, 1), (1, 1))#valid , ((1, 10, 108, 108), (20, 10, 102, 102), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_256 shape # test_lenet_256 shape
, ((2, 20, 124, 124), (30, 20, 9, 9), (1, 1), (1, 1), (1, 1))#valid , ((2, 20, 124, 124), (30, 20, 9, 9), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 2, 124, 124), (30, 2, 116, 116), (1, 1), (1, 1), (1, 1))#valid , ((20, 2, 124, 124), (30, 2, 116, 116), (1, 1), (1, 1), (1, 1)) # valid
, ((2, 1, 256, 256), (20, 1, 9, 9), (1, 1), (1, 1), (1, 1))#valid , ((2, 1, 256, 256), (20, 1, 9, 9), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 2, 256, 256), (20, 2, 248, 248), (1, 1), (1, 1), (1, 1))#valid , ((1, 2, 256, 256), (20, 2, 248, 248), (1, 1), (1, 1), (1, 1)) # valid
] ]
shapes_full = [ shapes_full = [
# test_lenet_28 shape # test_lenet_28 shape
((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#full ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # full
# test_lenet_32 shape # test_lenet_32 shape
, ((60, 30, 10, 10), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#full conv_full_patch_stack_padded' N=1 , ((60, 30, 10, 10), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # full conv_full_patch_stack_padded' N=1
# test_lenet_64 shape # test_lenet_64 shape
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#full conv_full_patch_stack_padded' N=3 , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # full conv_full_patch_stack_padded' N=3
# test_lenet_108 shape # test_lenet_108 shape
, ((10, 30, 45, 45), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#full 'conv_full_patch_stack_padded' N=9 , ((10, 30, 45, 45), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # full 'conv_full_patch_stack_padded' N=9
# test_lenet_256 shape # test_lenet_256 shape
, ((2, 30, 116, 116), (20, 30, 9, 9), (1, 1), (1, 1), (1, 1))#full conv_reference_full , ((2, 30, 116, 116), (20, 30, 9, 9), (1, 1), (1, 1), (1, 1)) # full conv_reference_full
] ]
version = [-1] version = [-1]
......
theano/sandbox/cuda/var.py
浏览文件 @ 3089deda
...@@ -89,7 +89,7 @@ class CudaNdarraySharedVariable(_operators, SharedVariable): ...@@ -89,7 +89,7 @@ class CudaNdarraySharedVariable(_operators, SharedVariable):
return self.container.value return self.container.value
else: else:
return copy.deepcopy(self.container.value) return copy.deepcopy(self.container.value)
else: #return an ndarray else: # return an ndarray
return numpy.asarray(self.container.value) return numpy.asarray(self.container.value)
def set_value(self, value, borrow=False): def set_value(self, value, borrow=False):
...@@ -129,7 +129,7 @@ class CudaNdarraySharedVariable(_operators, SharedVariable): ...@@ -129,7 +129,7 @@ class CudaNdarraySharedVariable(_operators, SharedVariable):
if not isinstance(value, numpy.ndarray): if not isinstance(value, numpy.ndarray):
# in case this is a cuda_ndarray, we copy it # in case this is a cuda_ndarray, we copy it
value = copy.deepcopy(value) value = copy.deepcopy(value)
self.container.value = value # this will copy a numpy ndarray self.container.value = value # this will copy a numpy ndarray
def __getitem__(self, *args): def __getitem__(self, *args):
# Defined to explicitly use the implementation from `_operators`, since # Defined to explicitly use the implementation from `_operators`, since
......
theano/sandbox/debug.py
浏览文件 @ 3089deda
...@@ -106,7 +106,7 @@ class DebugLinker(gof.WrapLinker): ...@@ -106,7 +106,7 @@ class DebugLinker(gof.WrapLinker):
fgraph = f.fgraph fgraph = f.fgraph
for r in fgraph.variables: for r in fgraph.variables:
if r.owner is None: if r.owner is None:
r.step = "value" # this will be overwritten if r is an input r.step = "value" # this will be overwritten if r is an input
else: else:
r.step = None r.step = None
r.value = None r.value = None
......
theano/sandbox/gpuarray/basic_ops.py
浏览文件 @ 3089deda
...@@ -696,7 +696,7 @@ class GpuAlloc(HideC, Alloc): ...@@ -696,7 +696,7 @@ class GpuAlloc(HideC, Alloc):
# If the output is a constant, it will have to be deepcopied # If the output is a constant, it will have to be deepcopied
# each time the function is called. So we do not fold. # each time the function is called. So we do not fold.
return False return False
elif (#The following ops work inplace of their input id 0. elif ( # The following ops work inplace of their input id 0.
client[1] == 0 and client[1] == 0 and
isinstance(client[0].op, ( isinstance(client[0].op, (
# Ops that will work inplace on the Alloc. So if they # Ops that will work inplace on the Alloc. So if they
......
theano/sandbox/gpuarray/elemwise.py
浏览文件 @ 3089deda
...@@ -1048,7 +1048,7 @@ class GpuCAReduceCuda(HideC, CAReduceDtype): ...@@ -1048,7 +1048,7 @@ class GpuCAReduceCuda(HideC, CAReduceDtype):
else: else:
assert isinstance(self.scalar_op, (scal.Maximum, assert isinstance(self.scalar_op, (scal.Maximum,
scal.Minimum)) scal.Minimum))
if self.pre_scalar_op: # TODO, multi_dtype! if self.pre_scalar_op: # TODO, multi_dtype!
#dtype = node.inputs[0].dtype #dtype = node.inputs[0].dtype
dtype = 'float32' dtype = 'float32'
......
theano/sandbox/gpuarray/tests/test_conv_cuda_ndarray.py
浏览文件 @ 3089deda
...@@ -372,28 +372,28 @@ def get_valid_shapes(): ...@@ -372,28 +372,28 @@ def get_valid_shapes():
, ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1))
, ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize , ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize
, ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image , ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image
, ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image, non-square kern , ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image, non-square kern
, ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim , ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim
, ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one , ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one
, ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1 , ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1
, ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights , ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights
, ((60, 20, 28, 28), (10, 20, 5, 5), (1, 1), (2, 2), (1, 1))#added a test case that fail from test_nnet.py.test_conv_nnet2 , ((60, 20, 28, 28), (10, 20, 5, 5), (1, 1), (2, 2), (1, 1)) # added a test case that fail from test_nnet.py.test_conv_nnet2
, ((10, 5, 28, 28), (10, 5, 5, 5), (1, 1), (2, 2), (1, 1))#test precedent but reduced that triger the error , ((10, 5, 28, 28), (10, 5, 5, 5), (1, 1), (2, 2), (1, 1)) # test precedent but reduced that triger the error
# Test more than maxThreadsDim0 # Test more than maxThreadsDim0
, ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 13, 1050), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
, ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1)) , ((2, 4, 1050, 13), (3, 4, 10, 11), (1, 1), (1, 1), (1, 1))
] ]
shapes += [ ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers shapes += [ ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 1 layers
, ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers , ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 2 layers
, ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 1 full , ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 1 full
, ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid , ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 2 valid
# , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
, ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers , ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 1 layers
, ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers , ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 2 layers
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 full , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 full
# , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1 # , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1
# , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2 # , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2
] ]
...@@ -635,30 +635,30 @@ def test_full(): ...@@ -635,30 +635,30 @@ def test_full():
, ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 3), (1, 1), (1, 1), (1, 1))
, ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 10, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1)) , ((4, 1, 20, 10), (1, 1, 2, 10), (1, 1), (1, 1), (1, 1))
, ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize , ((3, 2, 8, 8), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize
, ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image , ((3, 2, 8, 6), (4, 2, 4, 4), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image
, ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize, non-square image, non-square kern , ((3, 2, 8, 6), (4, 2, 4, 3), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize, non-square image, non-square kern
, ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) #stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim , ((3, 2, 8, 6), (4, 2, 4, 6), (1, 1), (1, 1), (1, 1)) # stack, nkern, bsize ,non-square image, non-square kern, kernsize==imgsize on one dim
, ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one , ((16, 5, 64, 64), (8, 5, 8, 8), (1, 1), (1, 1), (1, 1)) # a big one
, ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1 , ((16, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # MNIST LeNET layer 1
, ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights , ((20, 16, 32, 32), (1, 16, 28, 28), (1, 1), (1, 1), (1, 1)) # layer 1 backprop to weights
# other test # other test
, ((3, 1, 1, 1), (2, 1, 5, 3), (1, 1), (1, 1), (1, 1))#kernel bigger then image , ((3, 1, 1, 1), (2, 1, 5, 3), (1, 1), (1, 1), (1, 1)) # kernel bigger then image
, ((3, 2, 1, 1), (4, 2, 1, 1), (1, 1), (1, 1), (1, 1)) , ((3, 2, 1, 1), (4, 2, 1, 1), (1, 1), (1, 1), (1, 1))
, ((3, 2, 4, 4), (4, 2, 2, 6), (1, 1), (1, 1), (1, 1)) , ((3, 2, 4, 4), (4, 2, 2, 6), (1, 1), (1, 1), (1, 1))
, ((3, 2, 4, 4), (4, 2, 8, 6), (1, 1), (1, 1), (1, 1))#kernel bigger then image , ((3, 2, 4, 4), (4, 2, 8, 6), (1, 1), (1, 1), (1, 1)) # kernel bigger then image
, ((4, 2, 10, 10), (3, 2, 2, 12), (1, 1), (1, 1), (1, 1)) , ((4, 2, 10, 10), (3, 2, 2, 12), (1, 1), (1, 1), (1, 1))
] ]
shapes += [ shapes += [
# ((60,1,28,28),(20,1,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers # ((60,1,28,28),(20,1,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 1 layers
# , ((60,20,12,12),(30,20,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers # , ((60,20,12,12),(30,20,5,5), (1, 1), (1, 1), (1, 1))#test_lenet_28 2 layers
((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 1 full ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # test_lenet_28 bprop 1 full
# , ((20,60,12,12),(30,60,8,8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((20,60,12,12),(30,60,8,8), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
# , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid # , ((1,60,28,28),(20,60,24,24), (1, 1), (1, 1), (1, 1))#test_lenet_28 bprop 2 valid
# , ((10,1,64,64),(20,1,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers # , ((10,1,64,64),(20,1,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 1 layers
# , ((10,20,29,29),(30,20,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers # , ((10,20,29,29),(30,20,7,7), (1, 1), (1, 1), (1, 1))#test_lenet_64 2 layers
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#test_lenet_64 full , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # test_lenet_64 full
# , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1 # , ((20,10,29,29),(30,10,23,23), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 1
# , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2 # , ((1,10,64,64),(20,10,58,58), (1, 1), (1, 1), (1, 1))#test_lenet_64 bprop 2
# Test more than maxThreadsDim0 # Test more than maxThreadsDim0
...@@ -780,43 +780,43 @@ def benchmark(): ...@@ -780,43 +780,43 @@ def benchmark():
shapes_valid = [ shapes_valid = [
# test_lenet_28 shape # test_lenet_28 shape
((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1))#valid ((20, 60, 12, 12), (30, 60, 8, 8), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 20, 12, 12), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 1, 28, 28), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 60, 28, 28), (20, 60, 24, 24), (1, 1), (1, 1), (1, 1))#valid , ((1, 60, 28, 28), (20, 60, 24, 24), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_32 shape # test_lenet_32 shape
, ((20, 60, 14, 14), (30, 60, 10, 10), (1, 1), (1, 1), (1, 1))#valid , ((20, 60, 14, 14), (30, 60, 10, 10), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 20, 14, 14), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 20, 14, 14), (30, 20, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((60, 1, 32, 32), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1))#valid , ((60, 1, 32, 32), (20, 1, 5, 5), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 60, 32, 32), (20, 60, 28, 28), (1, 1), (1, 1), (1, 1))#valid , ((1, 60, 32, 32), (20, 60, 28, 28), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_64 shape # test_lenet_64 shape
, ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 20, 29, 29), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 10, 29, 29), (30, 10, 23, 23), (1, 1), (1, 1), (1, 1))#valid , ((20, 10, 29, 29), (30, 10, 23, 23), (1, 1), (1, 1), (1, 1)) # valid
, ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 1, 64, 64), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 10, 64, 64), (20, 10, 58, 58), (1, 1), (1, 1), (1, 1))#valid , ((1, 10, 64, 64), (20, 10, 58, 58), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_108 shape # test_lenet_108 shape
, ((10, 20, 51, 51), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 20, 51, 51), (30, 20, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 10, 51, 51), (30, 10, 45, 45), (1, 1), (1, 1), (1, 1))#valid , ((20, 10, 51, 51), (30, 10, 45, 45), (1, 1), (1, 1), (1, 1)) # valid
, ((10, 1, 108, 108), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1))#valid , ((10, 1, 108, 108), (20, 1, 7, 7), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 10, 108, 108), (20, 10, 102, 102), (1, 1), (1, 1), (1, 1))#valid , ((1, 10, 108, 108), (20, 10, 102, 102), (1, 1), (1, 1), (1, 1)) # valid
# test_lenet_256 shape # test_lenet_256 shape
, ((2, 20, 124, 124), (30, 20, 9, 9), (1, 1), (1, 1), (1, 1))#valid , ((2, 20, 124, 124), (30, 20, 9, 9), (1, 1), (1, 1), (1, 1)) # valid
, ((20, 2, 124, 124), (30, 2, 116, 116), (1, 1), (1, 1), (1, 1))#valid , ((20, 2, 124, 124), (30, 2, 116, 116), (1, 1), (1, 1), (1, 1)) # valid
, ((2, 1, 256, 256), (20, 1, 9, 9), (1, 1), (1, 1), (1, 1))#valid , ((2, 1, 256, 256), (20, 1, 9, 9), (1, 1), (1, 1), (1, 1)) # valid
, ((1, 2, 256, 256), (20, 2, 248, 248), (1, 1), (1, 1), (1, 1))#valid , ((1, 2, 256, 256), (20, 2, 248, 248), (1, 1), (1, 1), (1, 1)) # valid
] ]
shapes_full = [ shapes_full = [
# test_lenet_28 shape # test_lenet_28 shape
((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#full ((60, 30, 8, 8), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # full
# test_lenet_32 shape # test_lenet_32 shape
, ((60, 30, 10, 10), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1))#full conv_full_patch_stack_padded' N=1 , ((60, 30, 10, 10), (20, 30, 5, 5), (1, 1), (1, 1), (1, 1)) # full conv_full_patch_stack_padded' N=1
# test_lenet_64 shape # test_lenet_64 shape
, ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#full conv_full_patch_stack_padded' N=3 , ((10, 30, 23, 23), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # full conv_full_patch_stack_padded' N=3
# test_lenet_108 shape # test_lenet_108 shape
, ((10, 30, 45, 45), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1))#full 'conv_full_patch_stack_padded' N=9 , ((10, 30, 45, 45), (20, 30, 7, 7), (1, 1), (1, 1), (1, 1)) # full 'conv_full_patch_stack_padded' N=9
# test_lenet_256 shape # test_lenet_256 shape
, ((2, 30, 116, 116), (20, 30, 9, 9), (1, 1), (1, 1), (1, 1))#full conv_reference_full , ((2, 30, 116, 116), (20, 30, 9, 9), (1, 1), (1, 1), (1, 1)) # full conv_reference_full
] ]
# shapes_valid=shapes_valid[-1:] # shapes_valid=shapes_valid[-1:]
......
theano/sandbox/gpuarray/tests/test_elemwise.py
浏览文件 @ 3089deda
...@@ -124,39 +124,39 @@ class test_GpuCAReduceCuda(test_GpuCAReduceCPY): ...@@ -124,39 +124,39 @@ class test_GpuCAReduceCuda(test_GpuCAReduceCPY):
((1, 2), (1,)), ((1, 2), (1,)),
((100, 3, 1300), [1]), ((100, 3, 1300), [1]),
((0,), [0]), ((5,), [0]), ((0,), [0]), ((5,), [0]),
((0, 0), [0, 1]), ((1, 0), [0, 1]), ((5, 4), [0, 1]), ((33, 31), [0, 1]), ((5, 4), [1]), ((5, 4), [0]),#need something bigger then 32 for some opt test. ((0, 0), [0, 1]), ((1, 0), [0, 1]), ((5, 4), [0, 1]), ((33, 31), [0, 1]), ((5, 4), [1]), ((5, 4), [0]), # need something bigger then 32 for some opt test.
((5, 4, 3), [0]), ((5, 4, 3), [1]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [1, 2]), ((5, 4, 3), [0, 1, 2]), ((5, 4, 3), [0]), ((5, 4, 3), [1]), ((5, 4, 3), [0, 1]), ((5, 4, 3), [2]), ((5, 4, 3), [1, 2]), ((5, 4, 3), [0, 1, 2]),
((0, 0, 0, 0), [0, 1, 2, 3]), ((0, 0, 0, 0), [0, 1, 2, 3]),
((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]), ((5, 4, 3, 20), [2, 3]), ((5, 4, 3, 2), [0, 1, 2, 3]), ((5, 4, 3, 2), [0, 2, 3]), ((5, 4, 3, 2), [1, 2, 3]),
# test shape bigger then 4096 on each dimension to make sure that we work correctly when we don't have enough thread/block in each dimensions # test shape bigger then 4096 on each dimension to make sure that we work correctly when we don't have enough thread/block in each dimensions
((4100, 3), [0]), ((3, 4101), [0]),#10 ((4100, 3), [0]), ((3, 4101), [0]), # 10
((1024, 33), [0]), ((33, 1024), [0]),#10 ((1024, 33), [0]), ((33, 1024), [0]), # 10
((1025, 33), [0]), ((33, 1025), [0]),#10 ((1025, 33), [0]), ((33, 1025), [0]), # 10
((4100, 3), [1]), ((3, 4101), [1]),#01 ((4100, 3), [1]), ((3, 4101), [1]), # 01
((1024, 33), [1]), ((33, 1024), [1]),#01 ((1024, 33), [1]), ((33, 1024), [1]), # 01
((1025, 33), [1]), ((33, 1025), [1]),#01 ((1025, 33), [1]), ((33, 1025), [1]), # 01
((4100, 3), [0, 1]), ((3, 4101), [0, 1]),#11 ((4100, 3), [0, 1]), ((3, 4101), [0, 1]), # 11
((1024, 33), [0, 1]), ((33, 1024), [0, 1]),#01 ((1024, 33), [0, 1]), ((33, 1024), [0, 1]), # 01
((1025, 33), [0, 1]), ((33, 1025), [0, 1]),#01 ((1025, 33), [0, 1]), ((33, 1025), [0, 1]), # 01
((4100, 4, 3), [0]), ((5, 4100, 3), [0]), ((5, 4, 4100), [0]), ((3, 65536, 1), [0]),#100 ((4100, 4, 3), [0]), ((5, 4100, 3), [0]), ((5, 4, 4100), [0]), ((3, 65536, 1), [0]), # 100
((4100, 4, 3), [1]), ((5, 4100, 3), [1]), ((5, 4, 4100), [1]),#010 ((4100, 4, 3), [1]), ((5, 4100, 3), [1]), ((5, 4, 4100), [1]), # 010
((4100, 4, 3), [2]), ((5, 4100, 3), [2]), ((5, 4, 4100), [2]),#001 ((4100, 4, 3), [2]), ((5, 4100, 3), [2]), ((5, 4, 4100), [2]), # 001
((4100, 4, 3), [0, 1]), ((5, 4100, 3), [0, 1]), ((5, 4, 4100), [0, 1]),#110 ((4100, 4, 3), [0, 1]), ((5, 4100, 3), [0, 1]), ((5, 4, 4100), [0, 1]), # 110
((4100, 4, 3), [1, 2]), ((5, 4100, 3), [1, 2]), ((5, 4, 4100), [1, 2]),#011 ((4100, 4, 3), [1, 2]), ((5, 4100, 3), [1, 2]), ((5, 4, 4100), [1, 2]), # 011
#((4100,4,3),[0,2]),((5,4100,3),[0,2]),((5,4,4100),[0,2]),#101 ##not implemented #((4100,4,3),[0,2]),((5,4100,3),[0,2]),((5,4,4100),[0,2]),#101 ##not implemented
((4100, 4, 3), [0, 1, 2]), ((5, 4100, 3), [0, 1, 2]), ((5, 4, 4100), [0, 1, 2]),#111 ((4100, 4, 3), [0, 1, 2]), ((5, 4100, 3), [0, 1, 2]), ((5, 4, 4100), [0, 1, 2]), # 111
((65, 4, 3), [0, 1, 2]), ((5, 65, 3), [0, 1, 2]), ((5, 4, 65), [0, 1, 2]),#111 ((65, 4, 3), [0, 1, 2]), ((5, 65, 3), [0, 1, 2]), ((5, 4, 65), [0, 1, 2]), # 111
((4100, 4, 3, 2), [2, 3]), ((4, 4100, 3, 2), [2, 3]), ((4, 3, 4100, 2), [2, 3]), ((4, 3, 2, 4100), [2, 3]),#0011 ((4100, 4, 3, 2), [2, 3]), ((4, 4100, 3, 2), [2, 3]), ((4, 3, 4100, 2), [2, 3]), ((4, 3, 2, 4100), [2, 3]), # 0011
((4100, 4, 3, 2), [1, 3]), ((4, 4100, 3, 2), [1, 3]), ((4, 3, 4100, 2), [1, 3]), ((4, 3, 2, 4100), [1, 3]),#0101 ((4100, 4, 3, 2), [1, 3]), ((4, 4100, 3, 2), [1, 3]), ((4, 3, 4100, 2), [1, 3]), ((4, 3, 2, 4100), [1, 3]), # 0101
((4100, 4, 3, 2), [0, 2, 3]), ((4, 4100, 3, 2), [0, 2, 3]), ((4, 3, 4100, 2), [0, 2, 3]),#((4,3,2,4100),[0,2,3]),#1011 ((4100, 4, 3, 2), [0, 2, 3]), ((4, 4100, 3, 2), [0, 2, 3]), ((4, 3, 4100, 2), [0, 2, 3]), # ((4,3,2,4100),[0,2,3]),#1011
((4100, 4, 3, 2), [1, 2, 3]), ((4, 4100, 3, 2), [1, 2, 3]), ((4, 3, 4100, 2), [1, 2, 3]), ((4, 3, 2, 4100), [1, 2, 3]),#0111 ((4100, 4, 3, 2), [1, 2, 3]), ((4, 4100, 3, 2), [1, 2, 3]), ((4, 3, 4100, 2), [1, 2, 3]), ((4, 3, 2, 4100), [1, 2, 3]), # 0111
((65, 4, 3, 2), [1, 2, 3]), ((4, 65, 3, 2), [1, 2, 3]), ((4, 3, 65, 2), [1, 2, 3]), ((4, 3, 2, 65), [1, 2, 3]),#0111 ((65, 4, 3, 2), [1, 2, 3]), ((4, 65, 3, 2), [1, 2, 3]), ((4, 3, 65, 2), [1, 2, 3]), ((4, 3, 2, 65), [1, 2, 3]), # 0111
((4100, 2, 3, 4), [0, 1, 2, 3]), ((2, 4100, 3, 4), [0, 1, 2, 3]), ((2, 3, 4100, 4), [0, 1, 2, 3]), ((2, 3, 4, 4100), [0, 1, 2, 3]), ((128, 1, 2, 3), [0, 1, 2, 3]),#1111 ((4100, 2, 3, 4), [0, 1, 2, 3]), ((2, 4100, 3, 4), [0, 1, 2, 3]), ((2, 3, 4100, 4), [0, 1, 2, 3]), ((2, 3, 4, 4100), [0, 1, 2, 3]), ((128, 1, 2, 3), [0, 1, 2, 3]), # 1111
# test pattern implemented by reshape # test pattern implemented by reshape
# Skip them as this test the op directly, not the optimization with reshape # Skip them as this test the op directly, not the optimization with reshape
......
theano/sandbox/linalg/ops.py
浏览文件 @ 3089deda
...@@ -324,7 +324,7 @@ def no_transpose_symmetric(node): ...@@ -324,7 +324,7 @@ def no_transpose_symmetric(node):
@register_stabilize @register_stabilize
@local_optimizer(None) # XXX: solve is defined later and can't be used here @local_optimizer(None) # XXX: solve is defined later and can't be used here
def psd_solve_with_chol(node): def psd_solve_with_chol(node):
if node.op == solve: if node.op == solve:
A, b = node.inputs # result is solution Ax=b A, b = node.inputs # result is solution Ax=b
...@@ -340,7 +340,7 @@ def psd_solve_with_chol(node): ...@@ -340,7 +340,7 @@ def psd_solve_with_chol(node):
@register_stabilize @register_stabilize
@register_specialize @register_specialize
@local_optimizer(None) # XXX: det is defined later and can't be used here @local_optimizer(None) # XXX: det is defined later and can't be used here
def local_det_chol(node): def local_det_chol(node):
""" """
If we have det(X) and there is already an L=cholesky(X) If we have det(X) and there is already an L=cholesky(X)
......
theano/sandbox/symbolic_module.py
浏览文件 @ 3089deda
...@@ -184,7 +184,7 @@ def compile(smod, initial_values=None): ...@@ -184,7 +184,7 @@ def compile(smod, initial_values=None):
print p print p
reflected[thing] = p reflected[thing] = p
else: else:
reflected[thing] = None #TODO: how to reflect derived resuls? reflected[thing] = None # TODO: how to reflect derived resuls?
elif issymbolicmethod(thing): elif issymbolicmethod(thing):
reflected[thing] = compiled_functions[thing] reflected[thing] = compiled_functions[thing]
else : else :
...@@ -332,12 +332,12 @@ if 0: ...@@ -332,12 +332,12 @@ if 0:
x=T.dmatrix(), #our points, one point per row x=T.dmatrix(), #our points, one point per row
w=T.dmatrix(), #first layer weights w=T.dmatrix(), #first layer weights
b=T.dvector(), #first layer bias b=T.dvector(), #first layer bias
**kwargs #other things from logistic_regression **kwargs # other things from logistic_regression
): ):
hid = T.tanh(T.dot(x, w) + b) hid = T.tanh(T.dot(x, w) + b)
if top_part: if top_part:
print 'top_part', top_part, 'kwargs', kwargs print 'top_part', top_part, 'kwargs', kwargs
top = top_part(x=hid, **kwargs) # SymbolicModule top = top_part(x=hid, **kwargs) # SymbolicModule
def params(): return top.params() + [w, b] def params(): return top.params() + [w, b]
else: else:
def params(): return [w, b] def params(): return [w, b]
...@@ -353,7 +353,7 @@ if 0: ...@@ -353,7 +353,7 @@ if 0:
if 0: if 0:
class SymbolicModule(object): class SymbolicModule(object):
name = "__no_name__" #name of this module name = "__no_name__" # name of this module
variable_table = {} #map strings (names) to Variables variable_table = {} #map strings (names) to Variables
method_table = {} #map strings to compilable functions method_table = {} #map strings to compilable functions
...@@ -385,8 +385,8 @@ if 0: ...@@ -385,8 +385,8 @@ if 0:
b=T.dvector(), #first layer bias b=T.dvector(), #first layer bias
v=T.dmatrix(), #second layer weights v=T.dmatrix(), #second layer weights
c=T.dvector(), #second layer bias c=T.dvector(), #second layer bias
step=T.dscalar(), #step size for gradient descent step=T.dscalar(), # step size for gradient descent
l2_coef=T.dscalar() #l2 regularization amount l2_coef=T.dscalar() # l2 regularization amount
): ):
"""Idea A: """Idea A:
""" """
......
theano/sandbox/test_theano_object.py
浏览文件 @ 3089deda
...@@ -19,7 +19,7 @@ class MyModule(TheanoObject): ...@@ -19,7 +19,7 @@ class MyModule(TheanoObject):
super(MyModule, self).__init__() super(MyModule, self).__init__()
self.a = self.symbolic_member(2) self.a = self.symbolic_member(2)
self.b = self.symbolic_member(3) self.b = self.symbolic_member(3)
self.c = 100 #a constant self.c = 100 # a constant
self.d = [self.symbolic_member(5), self.symbolic_member(6)] self.d = [self.symbolic_member(5), self.symbolic_member(6)]
self.e = ['a', self.symbolic_member(6)] self.e = ['a', self.symbolic_member(6)]
...@@ -46,11 +46,11 @@ def test_outputs(): ...@@ -46,11 +46,11 @@ def test_outputs():
assert MM.add(5) == 12 assert MM.add(5) == 12
assert MM.b.get() == 4 assert MM.b.get() == 4
MM.sub(3) MM.sub(3)
assert MM.b.get() == 1 #test get() assert MM.b.get() == 1 # test get()
assert MM.add(5) == 9 #test that b's container is shared between add and sub assert MM.add(5) == 9 # test that b's container is shared between add and sub
MM.b.set(2) #test set MM.b.set(2) # test set
assert MM.b.get() == 2 #test get() assert MM.b.get() == 2 # test get()
assert MM.add(5) == 10 #test that b's container is shared between add and sub assert MM.add(5) == 10 # test that b's container is shared between add and sub
@run(True) @run(True)
def test_submodule(): def test_submodule():
...@@ -59,7 +59,7 @@ def test_submodule(): ...@@ -59,7 +59,7 @@ def test_submodule():
assert MM.add(5) == 8 assert MM.add(5) == 8
MM.submodule.sub(7) MM.submodule.sub(7)
assert MM.submodule.b.get() == -3 assert MM.submodule.b.get() == -3
assert MM.use_submodule(0) == -2 #self.a is 1 + self.submodule.b is -3 assert MM.use_submodule(0) == -2 # self.a is 1 + self.submodule.b is -3
@run(False) @run(False)
......
theano/scalar/basic.py
浏览文件 @ 3089deda
...@@ -3162,7 +3162,7 @@ class Composite(ScalarOp): ...@@ -3162,7 +3162,7 @@ class Composite(ScalarOp):
res = theano.compile.rebuild_collect_shared( res = theano.compile.rebuild_collect_shared(
inputs=inputs, inputs=inputs,
outputs=outputs[0].owner.inputs, outputs=outputs[0].owner.inputs,
copy_inputs_over=False) # Clone also the inputs copy_inputs_over=False) # Clone also the inputs
# 2. We continue this partial clone with the graph in # 2. We continue this partial clone with the graph in
# the inner Composite # the inner Composite
res2 = theano.compile.rebuild_collect_shared( res2 = theano.compile.rebuild_collect_shared(
......
theano/scan_module/scan_op.py
浏览文件 @ 3089deda
...@@ -1506,7 +1506,7 @@ class Scan(PureOp): ...@@ -1506,7 +1506,7 @@ class Scan(PureOp):
""" """
output = [] output = []
outer_inp_idx = 1 # First outer input is timestep index, skip it outer_inp_idx = 1 # First outer input is timestep index, skip it
# Handle sequences inputs # Handle sequences inputs
for i in range(self.info['n_seqs']): for i in range(self.info['n_seqs']):
......
theano/scan_module/scan_opt.py
浏览文件 @ 3089deda
...@@ -692,7 +692,7 @@ class PushOutScanOutput(gof.Optimizer): ...@@ -692,7 +692,7 @@ class PushOutScanOutput(gof.Optimizer):
outer_dot_inputs = [outer_vector_input, outer_dot_inputs = [outer_vector_input,
outer_matrix_input.transpose()] outer_matrix_input.transpose()]
outer_dot_output = theano.tensor.dot(*outer_dot_inputs) outer_dot_output = theano.tensor.dot(*outer_dot_inputs)
else: # idx_matrix_input == 1 else: # idx_matrix_input == 1
outer_dot_inputs = [outer_vector_input, outer_dot_inputs = [outer_vector_input,
outer_matrix_input] outer_matrix_input]
outer_dot_output = theano.tensor.dot(*outer_dot_inputs) outer_dot_output = theano.tensor.dot(*outer_dot_inputs)
...@@ -725,7 +725,7 @@ class PushOutScanOutput(gof.Optimizer): ...@@ -725,7 +725,7 @@ class PushOutScanOutput(gof.Optimizer):
sitsot_in_idx = nd.inputs.index(args.inner_in_sit_sot[sitsot_idx]) sitsot_in_idx = nd.inputs.index(args.inner_in_sit_sot[sitsot_idx])
dot_in_idx = 1 - sitsot_in_idx # 0 if sitsot_in_idx==1, dot_in_idx = 1 - sitsot_in_idx # 0 if sitsot_in_idx==1,
# 1 if sitsot_in_idx==0 # 1 if sitsot_in_idx==0
dot_input = nd.inputs[dot_in_idx] dot_input = nd.inputs[dot_in_idx]
......
theano/sparse/sandbox/test_sp.py
浏览文件 @ 3089deda
...@@ -140,7 +140,7 @@ class TestSP(unittest.TestCase): ...@@ -140,7 +140,7 @@ class TestSP(unittest.TestCase):
bsize = 10 # batch size bsize = 10 # batch size
imshp = (8, 8) imshp = (8, 8)
kshp = (5, 5) kshp = (5, 5)
nkern = 1 # per output pixel nkern = 1 # per output pixel
ssizes = ((1, 1), (2, 2)) ssizes = ((1, 1), (2, 2))
convmodes = ('full', 'valid',) convmodes = ('full', 'valid',)
...@@ -182,7 +182,7 @@ class TestSP(unittest.TestCase): ...@@ -182,7 +182,7 @@ class TestSP(unittest.TestCase):
patch = numpy.zeros((kshp[0], kshp[1])) patch = numpy.zeros((kshp[0], kshp[1]))
for b in xrange(bsize): for b in xrange(bsize):
for k in xrange(nkern): for k in xrange(nkern):
pixi = 0 # pixel index in raster order pixi = 0 # pixel index in raster order
for j in xrange(outshp[1]): for j in xrange(outshp[1]):
for i in xrange(outshp[2]): for i in xrange(outshp[2]):
n = j * ss[0] n = j * ss[0]
...@@ -225,7 +225,7 @@ class TestSP(unittest.TestCase): ...@@ -225,7 +225,7 @@ class TestSP(unittest.TestCase):
bsize = 10 # batch size bsize = 10 # batch size
imshp = (5, 5) imshp = (5, 5)
kshp = ((3, 3), (2, 2)) kshp = ((3, 3), (2, 2))
nkerns = (10, 20) # per output pixel nkerns = (10, 20) # per output pixel
ssizes = ((1, 1), (2, 2)) ssizes = ((1, 1), (2, 2))
convmodes = ('full', 'valid',) convmodes = ('full', 'valid',)
...@@ -263,9 +263,9 @@ class TestSP(unittest.TestCase): ...@@ -263,9 +263,9 @@ class TestSP(unittest.TestCase):
bsize = 10 # batch size bsize = 10 # batch size
imshp = (5, 5) imshp = (5, 5)
kshp = ((3, 3), (2, 2)) kshp = ((3, 3), (2, 2))
nkerns = (3, 6) # per output pixel nkerns = (3, 6) # per output pixel
ssizes = (((1, 1), (2, 2)),) ssizes = (((1, 1), (2, 2)),)
convmodes = ('full',)#'valid',) convmodes = ('full',) # 'valid',)
# symbolic stuff # symbolic stuff
kerns = [tensor.dmatrix(), tensor.dmatrix()] kerns = [tensor.dmatrix(), tensor.dmatrix()]
...@@ -338,7 +338,7 @@ class TestSP(unittest.TestCase): ...@@ -338,7 +338,7 @@ class TestSP(unittest.TestCase):
def test_CSMGrad(self): def test_CSMGrad(self):
imshp = (3, 3) imshp = (3, 3)
nkern = 1 # per output pixel nkern = 1 # per output pixel
kshp = (2, 2) kshp = (2, 2)
#ssizes = ((1,1),(2,2)) #ssizes = ((1,1),(2,2))
ssizes = ((1, 1),) ssizes = ((1, 1),)
......
theano/tensor/elemwise_cgen.py
浏览文件 @ 3089deda
...@@ -8,7 +8,7 @@ def make_declare(loop_orders, dtypes, sub): ...@@ -8,7 +8,7 @@ def make_declare(loop_orders, dtypes, sub):
decl = "" decl = ""
for i, (loop_order, dtype) in enumerate(zip(loop_orders, dtypes)): for i, (loop_order, dtype) in enumerate(zip(loop_orders, dtypes)):
var = sub['lv%i' % i] # input name corresponding to ith loop variable var = sub['lv%i' % i] # input name corresponding to ith loop variable
# we declare an iteration variable # we declare an iteration variable
# and an integer for the number of dimensions # and an integer for the number of dimensions
decl += """ decl += """
...@@ -229,7 +229,7 @@ def make_loop(loop_orders, dtypes, loop_tasks, sub, openmp=None): ...@@ -229,7 +229,7 @@ def make_loop(loop_orders, dtypes, loop_tasks, sub, openmp=None):
preloops.setdefault(j, "") preloops.setdefault(j, "")
preloops[j] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub preloops[j] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub
break break
else: # all broadcastable else: # all broadcastable
preloops.setdefault(0, "") preloops.setdefault(0, "")
preloops[0] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub preloops[0] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub
...@@ -512,7 +512,7 @@ def make_loop_careduce(loop_orders, dtypes, loop_tasks, sub): ...@@ -512,7 +512,7 @@ def make_loop_careduce(loop_orders, dtypes, loop_tasks, sub):
preloops.setdefault(j, "") preloops.setdefault(j, "")
preloops[j] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub preloops[j] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub
break break
else: # all broadcastable else: # all broadcastable
preloops.setdefault(0, "") preloops.setdefault(0, "")
preloops[0] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub preloops[0] += ("%%(lv%(i)s)s_iter = (%(dtype)s*)(PyArray_DATA(%%(lv%(i)s)s));\n" % locals()) % sub
......
theano/tensor/nnet/tests/speed_test_conv.py
浏览文件 @ 3089deda
...@@ -89,9 +89,9 @@ def exec_multilayer_conv_nnet_old(conv_mode, ss, bsize, imshp, kshps, nkerns, ...@@ -89,9 +89,9 @@ def exec_multilayer_conv_nnet_old(conv_mode, ss, bsize, imshp, kshps, nkerns,
from scipy.signal.signaltools import _valfrommode, _bvalfromboundary from scipy.signal.signaltools import _valfrommode, _bvalfromboundary
val = _valfrommode(conv_mode) val = _valfrommode(conv_mode)
bval = _bvalfromboundary('fill') bval = _bvalfromboundary('fill')
for b in range(bsize): # loop over batches for b in range(bsize): # loop over batches
for n in range(nkern): # loop over filters for n in range(nkern): # loop over filters
for i in range(imshp[0]): # loop over input feature maps for i in range(imshp[0]): # loop over input feature maps
outval[b, n, ...] += _convolve2d(\ outval[b, n, ...] += _convolve2d(\
imgval[b, i, ...], w_flip[n, i, ...], 1, val, bval, 0)[0::ss[0], 0::ss[1]] imgval[b, i, ...], w_flip[n, i, ...], 1, val, bval, 0)[0::ss[0], 0::ss[1]]
ntot += time.time() - time1 ntot += time.time() - time1
...@@ -111,14 +111,14 @@ def exec_multilayer_conv_nnet_old(conv_mode, ss, bsize, imshp, kshps, nkerns, ...@@ -111,14 +111,14 @@ def exec_multilayer_conv_nnet_old(conv_mode, ss, bsize, imshp, kshps, nkerns,
time1 = time.time() time1 = time.time()
for i in range(repeat): for i in range(repeat):
hidval2_ = propup2(imgval, w_flip) hidval2_ = propup2(imgval, w_flip)
hidval2 = hidval2_#[:,:,0::ss[0],0::ss[1]] hidval2 = hidval2_ # [:,:,0::ss[0],0::ss[1]]
tctot += time.time() - time1 tctot += time.time() - time1
if conv_op_py: if conv_op_py:
time1 = time.time() time1 = time.time()
for i in range(repeat): for i in range(repeat):
hidval3_ = propup3(imgval, w_flip) hidval3_ = propup3(imgval, w_flip)
hidval3 = hidval3_#[:,:,0::ss[0],0::ss[1]] hidval3 = hidval3_ # [:,:,0::ss[0],0::ss[1]]
tpytot += time.time() - time1 tpytot += time.time() - time1
assert (N.abs(hidval2-hidval3)<1e-5).all() assert (N.abs(hidval2-hidval3)<1e-5).all()
else: else:
...@@ -183,7 +183,7 @@ def exec_multilayer_conv_nnet(conv_mode, ss, bsize, imshp, kshps, nkerns, ...@@ -183,7 +183,7 @@ def exec_multilayer_conv_nnet(conv_mode, ss, bsize, imshp, kshps, nkerns,
time1 = time.time() time1 = time.time()
for i in range(repeat): for i in range(repeat):
hidval2_ = propup2(imgval, w_flip) hidval2_ = propup2(imgval, w_flip)
hidval2 = hidval2_#[:,:,0::ss[0],0::ss[1]] hidval2 = hidval2_ # [:,:,0::ss[0],0::ss[1]]
tctot += time.time() - time1 tctot += time.time() - time1
imshp = tuple(outshp) imshp = tuple(outshp)
...@@ -197,22 +197,22 @@ def speed_multilayer_conv(): ...@@ -197,22 +197,22 @@ def speed_multilayer_conv():
# calculate the speed up of different combination of unroll # calculate the speed up of different combination of unroll
# put the paramter to the same you will try. # put the paramter to the same you will try.
validate=False# we don't validate the result to have it much faster! validate=False # we don't validate the result to have it much faster!
repeat = 3 repeat = 3
verbose=1 verbose=1
unroll_batch = [1, 2, 3, 4, 5, 6, 10]#15, 30, 60 always much slower unroll_batch = [1, 2, 3, 4, 5, 6, 10] # 15, 30, 60 always much slower
unroll_kern = [1, 2, 3, 4, 5, 6, 10]#15, 30, 60 always much slower unroll_kern = [1, 2, 3, 4, 5, 6, 10] # 15, 30, 60 always much slower
#unroll_batch = [1,4,5] #unroll_batch = [1,4,5]
#unroll_kern = [1,4,5] #unroll_kern = [1,4,5]
#unroll_batch = [1,4] #unroll_batch = [1,4]
#unroll_kern = [1,4] #unroll_kern = [1,4]
unroll_patch = [True, False] unroll_patch = [True, False]
bsize = 60 # batch size bsize = 60 # batch size
imshp_start = (1, 48, 48)#un square shape to test more corner case. imshp_start = (1, 48, 48) # un square shape to test more corner case.
kshps = ([11, 12],)#un square shape to test more corner case. kshps = ([11, 12],) # un square shape to test more corner case.
nkerns = [60] # per output pixel nkerns = [60] # per output pixel
ssizes = [(1, 1), ]#(1,1)]#(2,2) bugged ssizes = [(1, 1), ] # (1,1)]#(2,2) bugged
convmodes = ['valid', 'full'] convmodes = ['valid', 'full']
do_convolve2=False do_convolve2=False
a=T.dmatrix() a=T.dmatrix()
...@@ -246,9 +246,9 @@ def speed_multilayer_conv(): ...@@ -246,9 +246,9 @@ def speed_multilayer_conv():
if unroll_b==1 and unroll_k==1: if unroll_b==1 and unroll_k==1:
# print "unroll 1/1",tctot # print "unroll 1/1",tctot
worst=tctot worst=tctot
timing[n_b, n_k]=[tctot, tpytot, ntot]#[sum(tctot), sum(tpytot), sum(ntot)] timing[n_b, n_k]=[tctot, tpytot, ntot] # [sum(tctot), sum(tpytot), sum(ntot)]
if not t_: if not t_:
t=timing[:, :, 0, :]#We select only the c timing. t=timing[:, :, 0, :] # We select only the c timing.
else: else:
t=t_ t=t_
t=N.asarray(t) t=N.asarray(t)
......
theano/tensor/nnet/tests/test_nnet.py
浏览文件 @ 3089deda
...@@ -1137,7 +1137,7 @@ def test_argmax_pushdown(): ...@@ -1137,7 +1137,7 @@ def test_argmax_pushdown():
# print 'AFTER' # print 'AFTER'
# for node in fgraph.toposort(): # for node in fgraph.toposort():
# print node.op # print node.op
assert len(fgraph.toposort()) == 4 # an output_guard is second assert len(fgraph.toposort()) == 4 # an output_guard is second
assert isinstance(fgraph.toposort()[0].op, tensor.Elemwise) assert isinstance(fgraph.toposort()[0].op, tensor.Elemwise)
assert isinstance(fgraph.toposort()[1].op, Softmax) assert isinstance(fgraph.toposort()[1].op, Softmax)
assert isinstance(fgraph.toposort()[2].op, tensor.CAReduce) assert isinstance(fgraph.toposort()[2].op, tensor.CAReduce)
......
theano/tensor/signal/tests/test_downsample.py
浏览文件 @ 3089deda
...@@ -245,8 +245,8 @@ class TestDownsampleFactorMax(utt.InferShapeTester): ...@@ -245,8 +245,8 @@ class TestDownsampleFactorMax(utt.InferShapeTester):
stridesizes = [(2, 2), (2, 2), (1, 1), (1, 2), (2, 2)] stridesizes = [(2, 2), (2, 2), (1, 1), (1, 2), (2, 2)]
paddingsizes = [(2, 2), (1, 2), (2, 1), (0, 0), (1, 1)] paddingsizes = [(2, 2), (1, 2), (2, 1), (0, 0), (1, 1)]
imgsizes = [(5, 5), (5, 5), (5, 6), (6, 5), (5, 5)] imgsizes = [(5, 5), (5, 5), (5, 6), (6, 5), (5, 5)]
m = 4 # minibatch m = 4 # minibatch
c = 10 # channel size c = 10 # channel size
images = tensor.dtensor4() images = tensor.dtensor4()
for indx in numpy.arange(len(maxpoolsizes)): for indx in numpy.arange(len(maxpoolsizes)):
imgsize = imgsizes[indx] imgsize = imgsizes[indx]
......
theano/tensor/tests/mlp_test.py
浏览文件 @ 3089deda
...@@ -269,7 +269,7 @@ def test_mlp(): ...@@ -269,7 +269,7 @@ def test_mlp():
# allocate symbolic variables for the data # allocate symbolic variables for the data
index = T.lscalar() # index to a [mini]batch index = T.lscalar() # index to a [mini]batch
x = T.matrix('x') # the data is presented as rasterized images x = T.matrix('x') # the data is presented as rasterized images
y = T.ivector('y') # the labels are presented as 1D vector of y = T.ivector('y') # the labels are presented as 1D vector of
# [int] labels # [int] labels
rng = numpy.random.RandomState(1234) rng = numpy.random.RandomState(1234)
......
theano/tensor/tests/test_blas_scipy.py
浏览文件 @ 3089deda
...@@ -43,19 +43,19 @@ class TestScipyGer(TestCase, TestOptimizationMixin): ...@@ -43,19 +43,19 @@ class TestScipyGer(TestCase, TestOptimizationMixin):
f = self.function([self.A, self.x, self.y], f = self.function([self.A, self.x, self.y],
self.A + tensor.outer(self.x, self.y)) self.A + tensor.outer(self.x, self.y))
self.assertFunctionContains(f, ScipyGer(destructive=False)) self.assertFunctionContains(f, ScipyGer(destructive=False))
self.run_f(f) #DebugMode tests correctness self.run_f(f) # DebugMode tests correctness
def test_A_plus_scaled_outer(self): def test_A_plus_scaled_outer(self):
f = self.function([self.A, self.x, self.y], f = self.function([self.A, self.x, self.y],
self.A + 0.1 * tensor.outer(self.x, self.y)) self.A + 0.1 * tensor.outer(self.x, self.y))
self.assertFunctionContains(f, ScipyGer(destructive=False)) self.assertFunctionContains(f, ScipyGer(destructive=False))
self.run_f(f) #DebugMode tests correctness self.run_f(f) # DebugMode tests correctness
def test_scaled_A_plus_scaled_outer(self): def test_scaled_A_plus_scaled_outer(self):
f = self.function([self.A, self.x, self.y], f = self.function([self.A, self.x, self.y],
0.2 * self.A + 0.1 * tensor.outer(self.x, self.y)) 0.2 * self.A + 0.1 * tensor.outer(self.x, self.y))
self.assertFunctionContains(f, gemm_no_inplace) self.assertFunctionContains(f, gemm_no_inplace)
self.run_f(f) #DebugMode tests correctness self.run_f(f) # DebugMode tests correctness
class TestBlasStridesScipy(TestBlasStrides): class TestBlasStridesScipy(TestBlasStrides):
mode = theano.compile.get_default_mode() mode = theano.compile.get_default_mode()
......
theano/tensor/tests/test_mpi.py
浏览文件 @ 3089deda
...@@ -53,8 +53,8 @@ def test_mpi_send_wait_cmp(): ...@@ -53,8 +53,8 @@ def test_mpi_send_wait_cmp():
waitnode = y.owner waitnode = y.owner
sendnode = y.owner.inputs[0].owner sendnode = y.owner.inputs[0].owner
addnode = z.owner addnode = z.owner
assert mpi_send_wait_cmp(sendnode, addnode) < 0 # send happens first assert mpi_send_wait_cmp(sendnode, addnode) < 0 # send happens first
assert mpi_send_wait_cmp(waitnode, addnode) > 0 # wait happens last assert mpi_send_wait_cmp(waitnode, addnode) > 0 # wait happens last
def test_mpi_tag_ordering(): def test_mpi_tag_ordering():
x = recv((2, 2), 'float32', 1, 12) x = recv((2, 2), 'float32', 1, 12)
......
theano/tensor/tests/test_opt.py
浏览文件 @ 3089deda
...@@ -983,12 +983,12 @@ class test_fusion(unittest.TestCase): ...@@ -983,12 +983,12 @@ class test_fusion(unittest.TestCase):
(fx-theano.tensor.true_div(fy, fz), (fx, fy, fz), (fxv, (fx-theano.tensor.true_div(fy, fz), (fx, fy, fz), (fxv,
fyv, fzv), 1, fxv-(fyv/fzv), 'float32'), fyv, fzv), 1, fxv-(fyv/fzv), 'float32'),
(fx-theano.tensor.int_div(ix*100, iy*1000), (fx, ix, (fx-theano.tensor.int_div(ix*100, iy*1000), (fx, ix,
iy), (fxv, ixv, iyv), 1, fxv-((ixv*100)//(iyv*1000)), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}), #40 iy), (fxv, ixv, iyv), 1, fxv-((ixv*100)//(iyv*1000)), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}), # 40
(fx-(fy/2), (fx, fy), (fxv, fyv), 1, fxv-(fyv/2), 'float32'), (fx-(fy/2), (fx, fy), (fxv, fyv), 1, fxv-(fyv/2), 'float32'),
(fx-(fy%fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv%fzv), 'float32'), (fx-(fy%fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv%fzv), 'float32'),
(fx-(fy>fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv>fzv), 'float32'), (fx-(fy>fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv>fzv), 'float32'),
(fx-(fy>=fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv>=fzv), 'float32'), (fx-(fy>=fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv>=fzv), 'float32'),
(fx-(fy<fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv<fzv), 'float32'),#45 (fx-(fy<fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv<fzv), 'float32'), # 45
(fx-(fy<=fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv<=fzv), 'float32'), (fx-(fy<=fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(fyv<=fzv), 'float32'),
(fx-T.eq(fy, fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-( (fx-T.eq(fy, fz), (fx, fy, fz), (fxv, fyv, fzv), 1, fxv-(
fyv==fzv), 'float32'), fyv==fzv), 'float32'),
...@@ -997,7 +997,7 @@ class test_fusion(unittest.TestCase): ...@@ -997,7 +997,7 @@ class test_fusion(unittest.TestCase):
(fx-fy+tensor.tan(fz), (fx, fy, fz), (fxv, fyv, fzv), 1, (fx-fy+tensor.tan(fz), (fx, fy, fz), (fxv, fyv, fzv), 1,
fxv-fyv+numpy.tan(fzv), 'float32'), fxv-fyv+numpy.tan(fzv), 'float32'),
(fx-fy+tensor.tanh(fz), (fx, fy, fz), (fxv, fyv, fzv), 1, (fx-fy+tensor.tanh(fz), (fx, fy, fz), (fxv, fyv, fzv), 1,
fxv-fyv+numpy.tanh(fzv), 'float32'),#50 fxv-fyv+numpy.tanh(fzv), 'float32'), # 50
(fx-fy+tensor.sin(fz), (fx, fy, fz), (fxv, fyv, fzv), 1, (fx-fy+tensor.sin(fz), (fx, fy, fz), (fxv, fyv, fzv), 1,
fxv-fyv+numpy.sin(fzv), 'float32'), fxv-fyv+numpy.sin(fzv), 'float32'),
(fx-fy+tensor.sinh(fz), (fx, fy, fz), (fxv, fyv, fzv), 1, (fx-fy+tensor.sinh(fz), (fx, fy, fz), (fxv, fyv, fzv), 1,
...@@ -1007,7 +1007,7 @@ class test_fusion(unittest.TestCase): ...@@ -1007,7 +1007,7 @@ class test_fusion(unittest.TestCase):
(fx-fy+theano.tensor.sqrt(fz), (fx, fy, fz), (fxv, fyv, (fx-fy+theano.tensor.sqrt(fz), (fx, fy, fz), (fxv, fyv,
fzv), 1, fxv-fyv+numpy.sqrt(fzv), 'float32'), fzv), 1, fxv-fyv+numpy.sqrt(fzv), 'float32'),
(fx-fy+theano.tensor.inv(fz), (fx, fy, fz), (fxv, fyv, (fx-fy+theano.tensor.inv(fz), (fx, fy, fz), (fxv, fyv,
fzv), 1, fxv-fyv+(1/fzv), 'float32'),#55 fzv), 1, fxv-fyv+(1/fzv), 'float32'), # 55
(fx-fy+theano.tensor.neg(fz), (fx, fy, fz), (fxv, fyv, (fx-fy+theano.tensor.neg(fz), (fx, fy, fz), (fxv, fyv,
fzv), 1, fxv-fyv+(-fzv), 'float32'), fzv), 1, fxv-fyv+(-fzv), 'float32'),
(fx-fy+theano.tensor.round(fz), (fx, fy, fz), (fxv, fyv, (fx-fy+theano.tensor.round(fz), (fx, fy, fz), (fxv, fyv,
...@@ -1018,7 +1018,7 @@ class test_fusion(unittest.TestCase): ...@@ -1018,7 +1018,7 @@ class test_fusion(unittest.TestCase):
(fx-theano.tensor.or_(iy, iz), (fx, iy, iz), (fxv, iyv, (fx-theano.tensor.or_(iy, iz), (fx, iy, iz), (fxv, iyv,
izv), 1, fxv-(iyv|izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}), izv), 1, fxv-(iyv|izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}),
(fx-theano.tensor.xor(iy, iz), (fx, iy, iz), (fxv, iyv, (fx-theano.tensor.xor(iy, iz), (fx, iy, iz), (fxv, iyv,
izv), 1, fxv-(iyv^izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}),#60 izv), 1, fxv-(iyv^izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}), # 60
(fx-theano.tensor.and_(iy, iz), (fx, iy, iz), (fxv, iyv, (fx-theano.tensor.and_(iy, iz), (fx, iy, iz), (fxv, iyv,
izv), 1, fxv-(iyv&izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}), izv), 1, fxv-(iyv&izv), {'custom': 'float64', 'numpy+floatX': config.floatX, 'numpy': 'float64'}),
(fx-theano.tensor.invert(iy), (fx, iy), (fxv, iyv), 1, (fx-theano.tensor.invert(iy), (fx, iy), (fxv, iyv), 1,
...@@ -1029,9 +1029,9 @@ class test_fusion(unittest.TestCase): ...@@ -1029,9 +1029,9 @@ class test_fusion(unittest.TestCase):
(theano.tensor.pow(fx*fy+fz, fx*fy), (fx, fy, fz), (fxv, (theano.tensor.pow(fx*fy+fz, fx*fy), (fx, fy, fz), (fxv,
fyv, fzv), 1, numpy.power(fxv*fyv+fzv, fxv*fyv), 'float32'), fyv, fzv), 1, numpy.power(fxv*fyv+fzv, fxv*fyv), 'float32'),
(fv+fy**fz, (fv, fy, fz), (fvv, fyv, fzv), 2, fvv+fyv**fzv, (fv+fy**fz, (fv, fy, fz), (fvv, fyv, fzv), 2, fvv+fyv**fzv,
'float32'),#fused with a dimshuffle #65 'float32'), # fused with a dimshuffle #65
(fv-fy+tensor.tanh(fz), (fv, fy, fz), (fvv, fyv, fzv), 2, (fv-fy+tensor.tanh(fz), (fv, fy, fz), (fvv, fyv, fzv), 2,
fvv-fyv+numpy.tanh(fzv), 'float32'),#fused with a dimshuffle fvv-fyv+numpy.tanh(fzv), 'float32'), # fused with a dimshuffle
# Cases where the same input is reused many times. # Cases where the same input is reused many times.
(theano.tensor.mul(fx, fx, fx, fx), (fx,), (fxv,), 1, fxv* (theano.tensor.mul(fx, fx, fx, fx), (fx,), (fxv,), 1, fxv*
...@@ -1041,7 +1041,7 @@ class test_fusion(unittest.TestCase): ...@@ -1041,7 +1041,7 @@ class test_fusion(unittest.TestCase):
(theano.tensor.mul(fx, ftanx, ftanx, fx), (fx,), (fxv,), (theano.tensor.mul(fx, ftanx, ftanx, fx), (fx,), (fxv,),
1, fxv*numpy.tan(fxv)*numpy.tan(fxv)*fxv, 'float32'), 1, fxv*numpy.tan(fxv)*numpy.tan(fxv)*fxv, 'float32'),
(theano.tensor.mul(ftanx, ftanx, fx+fy), (fx, fy), (fxv, (theano.tensor.mul(ftanx, ftanx, fx+fy), (fx, fy), (fxv,
fyv), 1, numpy.tan(fxv)*numpy.tan(fxv)*(fxv+fyv), 'float32'), # 70 fyv), 1, numpy.tan(fxv)*numpy.tan(fxv)*(fxv+fyv), 'float32'), # 70
# Cases with different broadcast pattern. They should not # Cases with different broadcast pattern. They should not
# be merged as this would duplicate computation # be merged as this would duplicate computation
......
theano/tensor/tests/test_shared_randomstreams.py
浏览文件 @ 3089deda
...@@ -43,7 +43,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -43,7 +43,7 @@ class T_SharedRandomStreams(unittest.TestCase):
gn_val0 = gn() gn_val0 = gn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
# print fn_val0 # print fn_val0
numpy_val0 = rng.uniform(size=(2, 2)) numpy_val0 = rng.uniform(size=(2, 2))
...@@ -86,7 +86,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -86,7 +86,7 @@ class T_SharedRandomStreams(unittest.TestCase):
random.seed(utt.fetch_seed()) random.seed(utt.fetch_seed())
rng = numpy.random.RandomState() rng = numpy.random.RandomState()
rng.set_state(random[out.rng].get_state()) #tests getitem rng.set_state(random[out.rng].get_state()) # tests getitem
fn_val0 = fn() fn_val0 = fn()
fn_val1 = fn() fn_val1 = fn()
...@@ -143,7 +143,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -143,7 +143,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.uniform(-1, 1, size=(2, 2)) numpy_val0 = rng.uniform(-1, 1, size=(2, 2))
numpy_val1 = rng.uniform(-1, 1, size=(2, 2)) numpy_val1 = rng.uniform(-1, 1, size=(2, 2))
...@@ -160,7 +160,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -160,7 +160,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.normal(-1, 2, size=(2, 2)) numpy_val0 = rng.normal(-1, 2, size=(2, 2))
numpy_val1 = rng.normal(-1, 2, size=(2, 2)) numpy_val1 = rng.normal(-1, 2, size=(2, 2))
...@@ -176,7 +176,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -176,7 +176,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.random_integers(-5, 5, size=(20, 20)) numpy_val0 = rng.random_integers(-5, 5, size=(20, 20))
numpy_val1 = rng.random_integers(-5, 5, size=(20, 20)) numpy_val1 = rng.random_integers(-5, 5, size=(20, 20))
...@@ -198,7 +198,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -198,7 +198,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.choice(10, (11, 8), True, None) numpy_val0 = rng.choice(10, (11, 8), True, None)
numpy_val1 = rng.choice(10, (11, 8), True, None) numpy_val1 = rng.choice(10, (11, 8), True, None)
...@@ -215,7 +215,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -215,7 +215,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.poisson(lam=5, size=(11, 8)) numpy_val0 = rng.poisson(lam=5, size=(11, 8))
numpy_val1 = rng.poisson(lam=5, size=(11, 8)) numpy_val1 = rng.poisson(lam=5, size=(11, 8))
...@@ -232,7 +232,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -232,7 +232,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
# rng.permutation outputs one vector at a time, so we iterate. # rng.permutation outputs one vector at a time, so we iterate.
numpy_val0 = numpy.asarray([rng.permutation(10) for i in range(20)]) numpy_val0 = numpy.asarray([rng.permutation(10) for i in range(20)])
...@@ -251,7 +251,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -251,7 +251,7 @@ class T_SharedRandomStreams(unittest.TestCase):
fn_val1 = fn() fn_val1 = fn()
rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30) rng_seed = numpy.random.RandomState(utt.fetch_seed()).randint(2**30)
rng = numpy.random.RandomState(int(rng_seed)) #int() is for 32bit rng = numpy.random.RandomState(int(rng_seed)) # int() is for 32bit
numpy_val0 = rng.multinomial(1, [0.1]*10, size=(4, 4)) numpy_val0 = rng.multinomial(1, [0.1]*10, size=(4, 4))
numpy_val1 = rng.multinomial(1, [0.1]*10, size=(4, 4)) numpy_val1 = rng.multinomial(1, [0.1]*10, size=(4, 4))
...@@ -419,7 +419,7 @@ class T_SharedRandomStreams(unittest.TestCase): ...@@ -419,7 +419,7 @@ class T_SharedRandomStreams(unittest.TestCase):
assert numpy.allclose(val0, numpy_val0) assert numpy.allclose(val0, numpy_val0)
assert numpy.allclose(val1, numpy_val1) assert numpy.allclose(val1, numpy_val1)
for i in range(10): # every test has 50% chance of passing even with non-matching random states for i in range(10): # every test has 50% chance of passing even with non-matching random states
val2 = g() val2 = g()
numpy_val2 = multinomial_rng.multinomial(n=1, pvals=[.5, .5]) numpy_val2 = multinomial_rng.multinomial(n=1, pvals=[.5, .5])
assert numpy.all(val2 == numpy_val2) assert numpy.all(val2 == numpy_val2)
......
theano/tensor/tests/test_sharedvar.py
浏览文件 @ 3089deda
...@@ -97,7 +97,7 @@ def makeSharedTester(shared_constructor_, ...@@ -97,7 +97,7 @@ def makeSharedTester(shared_constructor_,
values_to_div = .5 values_to_div = .5
if self.op_by_matrix: if self.op_by_matrix:
values_to_div = self.internal_type(numpy.ones(x.shape, dtype=dtype)/2)#supported for cudandarray, but not ndarray. values_to_div = self.internal_type(numpy.ones(x.shape, dtype=dtype)/2) # supported for cudandarray, but not ndarray.
assert self.test_internal_type(values_to_div) assert self.test_internal_type(values_to_div)
x /= values_to_div x /= values_to_div
total_val_2 = total_func() total_val_2 = total_func()
...@@ -193,7 +193,7 @@ def makeSharedTester(shared_constructor_, ...@@ -193,7 +193,7 @@ def makeSharedTester(shared_constructor_,
# supported for cudandarray, but not ndarray. # supported for cudandarray, but not ndarray.
values_to_div = self.internal_type( values_to_div = self.internal_type(
numpy.ones(x.shape, dtype=dtype)/2) numpy.ones(x.shape, dtype=dtype)/2)
x /= values_to_div#supported by ndarray and CudaNdarray x /= values_to_div # supported by ndarray and CudaNdarray
# this is not required by the contract but it is a feature we can # this is not required by the contract but it is a feature we can
# implement for some type of SharedVariable. # implement for some type of SharedVariable.
...@@ -202,7 +202,7 @@ def makeSharedTester(shared_constructor_, ...@@ -202,7 +202,7 @@ def makeSharedTester(shared_constructor_,
x = x_shared.get_value(borrow=False, return_internal_type=True) x = x_shared.get_value(borrow=False, return_internal_type=True)
assert self.test_internal_type(x) assert self.test_internal_type(x)
assert x is not x_shared.container.value assert x is not x_shared.container.value
x /= values_to_div#supported by ndarray and CudaNdarray x /= values_to_div # supported by ndarray and CudaNdarray
# this is required by the contract # this is required by the contract
assert not numpy.allclose(self.ref_fct(x), total_func()) assert not numpy.allclose(self.ref_fct(x), total_func())
...@@ -251,7 +251,7 @@ def makeSharedTester(shared_constructor_, ...@@ -251,7 +251,7 @@ def makeSharedTester(shared_constructor_,
# test if that theano shared variable optimize set_value(borrow=True) # test if that theano shared variable optimize set_value(borrow=True)
get_x = x_shared.get_value(borrow=True) get_x = x_shared.get_value(borrow=True)
assert get_x is not x_orig#borrow=False to shared_constructor assert get_x is not x_orig # borrow=False to shared_constructor
get_x /= values_to_div get_x /= values_to_div
x_shared.set_value(get_x, borrow=True) x_shared.set_value(get_x, borrow=True)
x = x_shared.get_value(borrow=True) x = x_shared.get_value(borrow=True)
...@@ -263,10 +263,10 @@ def makeSharedTester(shared_constructor_, ...@@ -263,10 +263,10 @@ def makeSharedTester(shared_constructor_,
# test optimized get set value on the gpu(don't pass data to the cpu) # test optimized get set value on the gpu(don't pass data to the cpu)
get_x = x_shared.get_value(borrow=True, return_internal_type=True) get_x = x_shared.get_value(borrow=True, return_internal_type=True)
assert get_x is not x_orig#borrow=False to shared_constructor assert get_x is not x_orig # borrow=False to shared_constructor
assert self.test_internal_type(get_x) assert self.test_internal_type(get_x)
get_x /= values_to_div#supported by ndarray and CudaNdarray get_x /= values_to_div # supported by ndarray and CudaNdarray
assert self.test_internal_type(get_x) assert self.test_internal_type(get_x)
x_shared.set_value(get_x, borrow=True) x_shared.set_value(get_x, borrow=True)
x = x_shared.get_value(borrow=True, return_internal_type=True) x = x_shared.get_value(borrow=True, return_internal_type=True)
...@@ -316,7 +316,7 @@ def makeSharedTester(shared_constructor_, ...@@ -316,7 +316,7 @@ def makeSharedTester(shared_constructor_,
if dtype is None: if dtype is None:
dtype = theano.config.floatX dtype = theano.config.floatX
shp = (100/4, 1024)#100KB shp = (100/4, 1024) # 100KB
x = numpy.zeros(shp, dtype=dtype) x = numpy.zeros(shp, dtype=dtype)
x = self.cast_value(x) x = self.cast_value(x)
......
theano/tests/test_tutorial.py
浏览文件 @ 3089deda
...@@ -461,7 +461,7 @@ class T_extending(unittest.TestCase): ...@@ -461,7 +461,7 @@ class T_extending(unittest.TestCase):
# This should be needed for the EquilibriumOptimizer # This should be needed for the EquilibriumOptimizer
# but it isn't now # but it isn't now
# TODO: do this and explain it # TODO: do this and explain it
return [] # that's not what you should do return [] # that's not what you should do
local_simplify = LocalSimplify() local_simplify = LocalSimplify()
...@@ -775,8 +775,8 @@ class T_examples(unittest.TestCase): ...@@ -775,8 +775,8 @@ class T_examples(unittest.TestCase):
# Construct Theano expression graph # Construct Theano expression graph
p_1 = 1 / (1 + T.exp(-T.dot(x, w) - b)) # Probability that target = 1 p_1 = 1 / (1 + T.exp(-T.dot(x, w) - b)) # Probability that target = 1
prediction = p_1 > 0.5 # The prediction thresholded prediction = p_1 > 0.5 # The prediction thresholded
xent = -y * T.log(p_1) - (1-y) * T.log(1-p_1) # Cross-entropy loss function xent = -y * T.log(p_1) - (1-y) * T.log(1-p_1) # Cross-entropy loss function
cost = xent.mean() + 0.01 * (w ** 2).sum()# The cost to minimize cost = xent.mean() + 0.01 * (w ** 2).sum() # The cost to minimize
gw, gb = T.grad(cost, [w, b]) # Compute the gradient of the cost gw, gb = T.grad(cost, [w, b]) # Compute the gradient of the cost
# (we shall return to this in a # (we shall return to this in a
# following section of this tutorial) # following section of this tutorial)
...@@ -817,7 +817,7 @@ class T_aliasing(unittest.TestCase): ...@@ -817,7 +817,7 @@ class T_aliasing(unittest.TestCase):
s_false = theano.shared(np_array, borrow=False) s_false = theano.shared(np_array, borrow=False)
s_true = theano.shared(np_array, borrow=True) s_true = theano.shared(np_array, borrow=True)
np_array += 1 # now it is an array of 2.0 s np_array += 1 # now it is an array of 2.0 s
assert numpy.all(s_default.get_value() == array([1.0, 1.0])) assert numpy.all(s_default.get_value() == array([1.0, 1.0]))
assert numpy.all(s_false.get_value() == array([1.0, 1.0])) assert numpy.all(s_false.get_value() == array([1.0, 1.0]))
...@@ -831,7 +831,7 @@ class T_aliasing(unittest.TestCase): ...@@ -831,7 +831,7 @@ class T_aliasing(unittest.TestCase):
s = theano.shared(np_array) s = theano.shared(np_array)
v_false = s.get_value(borrow=False) # N.B. borrow default is False v_false = s.get_value(borrow=False) # N.B. borrow default is False
v_true = s.get_value(borrow=True) v_true = s.get_value(borrow=True)
v_internal = s.get_value(borrow=True, return_internal_type=True) v_internal = s.get_value(borrow=True, return_internal_type=True)
......
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论