Skip to content

P
pytensor
Unverified 提交 9f7a1b69 authored 作者: Brandon T. Willard's avatar Brandon T. Willard 提交者： GitHub
浏览文件
操作

Merge pull request #110 from brandonwillard/enforce-sane-test-values

Refactor test value framework so that test value validation is performed up-front.
上级 53486aae ea44b16d
隐藏空白字符变更
内嵌 并排
正在显示 包含 192 行增加 和 464 行删除
doc/library/misc/pkl_utils.txt
浏览文件 @ 9f7a1b69
......@@ -15,10 +15,6 @@
.. autoclass:: theano.misc.pkl_utils.StripPickler
.. autoclass:: theano.misc.pkl_utils.CompatUnpickler
.. seealso::
:ref:`tutorial_loadsave`
tests/gof/test_compute_test_value.py
浏览文件 @ 9f7a1b69
......@@ -167,14 +167,16 @@ class TestComputeTestValue:
@theano.change_flags(compute_test_value="raise")
def test_incorrect_type(self):
x = tt.fmatrix("x")
# Incorrect dtype (float64) for test_value
x.tag.test_value = np.random.rand(3, 4)
y = tt.dmatrix("y")
y.tag.test_value = np.random.rand(4, 5)
x = tt.vector("x")
with pytest.raises(TypeError):
tt.dot(x, y)
# Incorrect shape for test value
x.tag.test_value = np.empty((2, 2))
x = tt.fmatrix("x")
with pytest.raises(TypeError):
# Incorrect dtype (float64) for test value
x.tag.test_value = np.random.rand(3, 4)
@theano.change_flags(compute_test_value="raise")
def test_overided_function(self):
......
tests/gof/test_fg.py
浏览文件 @ 9f7a1b69
import os
import pickle
import pytest
import theano
from theano.compat import PY3
from theano.gof.fg import FunctionGraph
from theano import tensor as tt
from theano.gof.fg import FunctionGraph
class TestFunctionGraph:
......@@ -16,24 +11,3 @@ class TestFunctionGraph:
s = pickle.dumps(func)
pickle.loads(s)
@pytest.mark.skipif(
not theano.config.cxx, reason="G++ not available, so we need to skip this test."
)
@pytest.mark.slow
def test_node_outputs_not_used(self):
# In the past, we where removing some not used variable from
# fgraph.variables event if the apply had other output used in
# the graph. This caused a crash.
# This test run the pickle that reproduce this case.
with open(
os.path.join(os.path.dirname(__file__), "test_fg_old_crash.pkl"), "rb"
) as f:
from theano.misc.pkl_utils import CompatUnpickler
if PY3:
u = CompatUnpickler(f, encoding="latin1")
else:
u = CompatUnpickler(f)
d = u.load()
f = theano.function(**d)
tests/gof/test_op.py
浏览文件 @ 9f7a1b69
import numpy as np
import pytest
import theano
import theano.gof.op as op
import theano.tensor as tt
from six import string_types
from theano.gof.type import Type, Generic
from theano import scalar, shared
from theano.configparser import change_flags
from theano.gof.graph import Apply, Variable
import theano.tensor as T
from theano import scalar
from theano import shared
from theano.gof.type import Generic, Type
config = theano.config
Op = op.Op
......@@ -238,15 +238,15 @@ class TestMakeThunk:
__props__ = ()
itypes = [T.dmatrix]
otypes = [T.dmatrix]
itypes = [tt.dmatrix]
otypes = [tt.dmatrix]
def perform(self, node, inputs, outputs):
inp = inputs[0]
output = outputs[0]
output[0] = inp * 2
x_input = T.dmatrix("x_input")
x_input = tt.dmatrix("x_input")
f = theano.function([x_input], DoubleOp()(x_input))
inp = np.random.rand(5, 4)
out = f(inp)
......@@ -255,17 +255,17 @@ class TestMakeThunk:
def test_test_value_python_objects():
for x in ([0, 1, 2], 0, 0.5, 1):
assert (op.get_test_value(x) == x).all()
assert np.all(op.get_test_value(x) == x)
def test_test_value_ndarray():
x = np.zeros((5, 5))
v = op.get_test_value(x)
assert (v == x).all()
assert np.all(v == x)
def test_test_value_constant():
x = T.as_tensor_variable(np.zeros((5, 5)))
x = tt.as_tensor_variable(np.zeros((5, 5)))
v = op.get_test_value(x)
assert np.all(v == np.zeros((5, 5)))
......@@ -278,62 +278,37 @@ def test_test_value_shared():
assert np.all(v == np.zeros((5, 5)))
@change_flags(compute_test_value="raise")
def test_test_value_op():
try:
prev_value = config.compute_test_value
config.compute_test_value = "raise"
x = T.log(np.ones((5, 5)))
v = op.get_test_value(x)
assert np.allclose(v, np.zeros((5, 5)))
finally:
config.compute_test_value = prev_value
def test_get_debug_values_no_debugger():
"get_debug_values should return [] when debugger is off"
x = tt.log(np.ones((5, 5)))
v = op.get_test_value(x)
prev_value = config.compute_test_value
try:
config.compute_test_value = "off"
assert np.allclose(v, np.zeros((5, 5)))
x = T.vector()
for x_val in op.get_debug_values(x):
assert False
@change_flags(compute_test_value="off")
def test_get_debug_values_no_debugger():
"""Tests that `get_debug_values` returns `[]` when debugger is off."""
finally:
config.compute_test_value = prev_value
x = tt.vector()
assert op.get_debug_values(x) == []
@change_flags(compute_test_value="ignore")
def test_get_det_debug_values_ignore():
# get_debug_values should return [] when debugger is ignore
# and some values are missing
"""Tests that `get_debug_values` returns `[]` when debugger is set to "ignore" and some values are missing."""
prev_value = config.compute_test_value
try:
config.compute_test_value = "ignore"
x = T.vector()
for x_val in op.get_debug_values(x):
assert False
finally:
config.compute_test_value = prev_value
x = tt.vector()
assert op.get_debug_values(x) == []
def test_get_debug_values_success():
# tests that get_debug_value returns values when available
# (and the debugger is on)
"""Tests that `get_debug_value` returns values when available (and the debugger is on)."""
prev_value = config.compute_test_value
for mode in ["ignore", "warn", "raise"]:
try:
config.compute_test_value = mode
x = T.vector()
with change_flags(compute_test_value=mode):
x = tt.vector()
x.tag.test_value = np.zeros((4,), dtype=config.floatX)
y = np.zeros((5, 5))
......@@ -348,54 +323,11 @@ def test_get_debug_values_success():
assert iters == 1
finally:
config.compute_test_value = prev_value
@change_flags(compute_test_value="raise")
def test_get_debug_values_exc():
# tests that get_debug_value raises an exception when
# debugger is set to raise and a value is missing
prev_value = config.compute_test_value
try:
config.compute_test_value = "raise"
x = T.vector()
try:
for x_val in op.get_debug_values(x):
# this assert catches the case where we
# erroneously get a value returned
assert False
raised = False
except AttributeError:
raised = True
# this assert catches the case where we got []
# returned, and possibly issued a warning,
# rather than raising an exception
assert raised
"""Tests that `get_debug_value` raises an exception when debugger is set to raise and a value is missing."""
finally:
config.compute_test_value = prev_value
def test_debug_error_message():
# tests that debug_error_message raises an
# exception when it should.
prev_value = config.compute_test_value
for mode in ["ignore", "raise"]:
try:
config.compute_test_value = mode
try:
op.debug_error_message("msg")
raised = False
except ValueError:
raised = True
assert raised
finally:
config.compute_test_value = prev_value
with pytest.raises(AttributeError):
x = tt.vector()
assert op.get_debug_values(x) == []
tests/gpuarray/test_multinomial.py
浏览文件 @ 9f7a1b69
......@@ -7,9 +7,10 @@ import theano
import tests.unittest_tools as utt
from pickle import Unpickler
from theano import config, function, tensor
from theano.compat import PY3
from theano.misc.pkl_utils import CompatUnpickler
from theano.sandbox import multinomial
from theano.sandbox.rng_mrg import MRG_RandomStreams as RandomStreams
from theano.gpuarray.multinomial import (
......@@ -384,6 +385,6 @@ def test_unpickle_legacy_op():
if not PY3:
with open(os.path.join(testfile_dir, fname), "r") as fp:
u = CompatUnpickler(fp)
u = Unpickler(fp)
m = u.load()
assert isinstance(m, GPUAChoiceFromUniform)
tests/gpuarray/test_pickle.py
浏览文件 @ 9f7a1b69
......@@ -13,9 +13,9 @@ import pytest
import numpy as np
from pickle import Unpickler
from theano import config
from theano.compat import PY3
from theano.misc.pkl_utils import CompatUnpickler
from theano.gpuarray.type import ContextNotDefined
......@@ -37,10 +37,7 @@ def test_unpickle_gpuarray_as_numpy_ndarray_flag1():
fname = "GpuArray.pkl"
with open(os.path.join(testfile_dir, fname), "rb") as fp:
if PY3:
u = CompatUnpickler(fp, encoding="latin1")
else:
u = CompatUnpickler(fp)
u = Unpickler(fp, encoding="latin1")
with pytest.raises((ImportError, ContextNotDefined)):
u.load()
finally:
......@@ -56,10 +53,7 @@ def test_unpickle_gpuarray_as_numpy_ndarray_flag2():
fname = "GpuArray.pkl"
with open(os.path.join(testfile_dir, fname), "rb") as fp:
if PY3:
u = CompatUnpickler(fp, encoding="latin1")
else:
u = CompatUnpickler(fp)
u = Unpickler(fp, encoding="latin1")
try:
mat = u.load()
except ImportError:
......
tests/gpuarray/test_type.py
浏览文件 @ 9f7a1b69
......@@ -5,10 +5,10 @@ import theano
pygpu = pytest.importorskip("pygpu")
from theano.compat import PY3
from pickle import Unpickler
from theano import config
from theano.compile import DeepCopyOp, Rebroadcast, ViewOp
from theano.misc.pkl_utils import CompatUnpickler
from theano.gpuarray.type import GpuArrayType, gpuarray_shared_constructor
from tests.gpuarray.config import test_ctx_name
......@@ -122,10 +122,7 @@ def test_unpickle_gpuarray_as_numpy_ndarray_flag0():
fname = "GpuArray.pkl"
with open(os.path.join(testfile_dir, fname), "rb") as fp:
if PY3:
u = CompatUnpickler(fp, encoding="latin1")
else:
u = CompatUnpickler(fp)
u = Unpickler(fp, encoding="latin1")
mat = u.load()
assert isinstance(mat, pygpu.gpuarray.GpuArray)
assert np.asarray(mat)[0] == -42.0
......
tests/sandbox/test_multinomial.py
浏览文件 @ 9f7a1b69
import os
import sys
import numpy as np
import theano
import tests.unittest_tools as utt
from theano import config, function, tensor
from theano.sandbox import multinomial
from theano.compat import PY3
from theano.misc.pkl_utils import CompatUnpickler
def test_n_samples_1():
......@@ -51,40 +45,6 @@ def test_n_samples_2():
assert res.sum() == i
def test_n_samples_compatibility():
# This test checks if the new change to MultinomialFromUniform is still compatible
# with old interface. Here I will load a graph created (using the old interface) as follows:
# RandomStreams = theano.sandbox.rng_mrg.MRG_RandomStreams
# th_rng = RandomStreams(12345)
# X = T.matrix('X')
# pvals = T.exp(X)
# pvals = pvals / pvals.sum(axis=1, keepdims=True)
# samples = th_rng.multinomial(pvals=pvals)
# pickle.dump([X, samples], open("multinomial_test_graph.pkl", "w"))
folder = os.path.dirname(os.path.abspath(__file__))
with open(os.path.join(folder, "multinomial_test_graph.pkl"), "rb") as pkl_file:
if PY3:
u = CompatUnpickler(pkl_file, encoding="latin1")
else:
u = CompatUnpickler(pkl_file)
try:
X, samples = u.load()
except ImportError:
# Windows sometimes fail with nonsensical errors like:
# ImportError: No module named type
# ImportError: No module named copy_reg
# when "type" and "copy_reg" are builtin modules.
if sys.platform == "win32":
exc_type, exc_value, exc_trace = sys.exc_info()
raise
raise
f = theano.function([X], samples)
res = f(np.random.randn(20, 10))
assert np.all(res.sum(axis=1) == 1)
def test_multinomial_0():
# This tests the MultinomialFromUniform Op directly, not going through the
# multinomial() call in GPU random generation.
......
tests/sandbox/test_multinomial_wo_replacement.py
浏览文件 @ 9f7a1b69
import numpy as np
import pytest
import os
from theano import config, function, tensor
from theano.compat import PY3
from theano.misc.pkl_utils import CompatUnpickler
from theano.sandbox import multinomial
from theano.sandbox.rng_mrg import MRG_RandomStreams as RandomStreams
......@@ -214,14 +212,3 @@ class TestFunction:
avg_pvals /= avg_pvals.sum()
avg_diff = np.mean(abs(avg_pvals - pvals))
assert avg_diff < mean_rtol
def test_unpickle_legacy_op(self):
testfile_dir = os.path.dirname(os.path.realpath(__file__))
fname = "test_sandbox_multinomial_wo_replacement.pkl"
if not PY3:
with open(os.path.join(testfile_dir, fname), "r") as fp:
u = CompatUnpickler(fp)
m = u.load()
print(m)
assert isinstance(m, multinomial.ChoiceFromUniform)
tests/unittest_tools.py
浏览文件 @ 9f7a1b69
......@@ -255,7 +255,7 @@ class InferShapeTester:
else:
shp = inp.shape
if len(set(shp)) != len(shp):
_logger.warn(
_logger.warning(
"While testing shape inference for %r, we received an"
" input with a shape that has some repeated values: %r"
", like a square matrix. This makes it impossible to"
......
theano/compile/debugmode.py
浏览文件 @ 9f7a1b69
......@@ -1437,7 +1437,7 @@ def _check_preallocated_output(
fn_attr_name = ops_with_inner_function[type(node.op)]
fn = getattr(node.op, fn_attr_name, None)
if not fn or not hasattr(fn, "maker") or not hasattr(fn.maker, "mode"):
_logger.warn(
_logger.warning(
"Expected theano function not found in %s.%s", node.op, fn_attr_name
)
else:
......@@ -1482,7 +1482,7 @@ def _check_preallocated_output(
if not out_map:
# Map is empty, there is no need to execute thunk() again
_logger.warn("%s: out_map is empty", name)
_logger.warning("%s: out_map is empty", name)
continue
# Copy the inputs over, if they were marked as destroyed or viewed
......@@ -1904,7 +1904,7 @@ class _Linker(gof.link.LocalLinker):
thunks_py.append(None)
if not self.maker.mode.check_c_code and thunks_py[-1] is None:
_logger.warn(
_logger.warning(
"Op %s doesn't have a perform, "
"forcing check of the C code" % node.op
)
......@@ -1921,7 +1921,7 @@ class _Linker(gof.link.LocalLinker):
elif thunks_c[-1] is None:
thunks_c[-1] = thunk_other
else:
_logger.warn(
_logger.warning(
"We won't check the perform function "
"of node '%s' but we will check its "
"make_thunk function" % node
......
theano/gof/cmodule.py
浏览文件 @ 9f7a1b69
......@@ -2055,7 +2055,7 @@ class GCC_compiler(Compiler):
and "clang-omp++" not in theano.config.cxx
and "icpc" not in theano.config.cxx
):
_logger.warn(
_logger.warning(
"OPTIMIZATION WARNING: your Theano flag `cxx` seems not to be"
" the g++ compiler. So we disable the compiler optimization"
" specific to g++ that tell to compile for a specific CPU."
......@@ -2124,7 +2124,7 @@ class GCC_compiler(Compiler):
)
else:
reported_lines = native_lines
_logger.warn(
_logger.warning(
"OPTIMIZATION WARNING: Theano was not able to find the"
" g++ parameters that tune the compilation to your "
" specific CPU. This can slow down the execution of Theano"
......@@ -2137,7 +2137,7 @@ class GCC_compiler(Compiler):
default_lines = get_lines("%s -E -v -" % theano.config.cxx)
_logger.info("g++ default lines: %s", default_lines)
if len(default_lines) < 1:
_logger.warn(
_logger.warning(
"OPTIMIZATION WARNING: Theano was not able to find the"
" default g++ parameters. This is needed to tune"
" the compilation to your specific"
......
theano/gof/compilelock.py
浏览文件 @ 9f7a1b69
......@@ -349,7 +349,7 @@ def refresh_lock(lock_file):
# This way, only 1 test would fail.
while get_lock.n_lock > 0:
release_lock()
_logger.warn(
_logger.warning(
"Refreshing lock failed, we release the"
" lock before raising again the exception"
)
......
theano/gof/graph.py
浏览文件 @ 9f7a1b69
......@@ -92,7 +92,7 @@ class Apply(Node):
def __init__(self, op, inputs, outputs):
self.op = op
self.inputs = []
self.tag = utils.scratchpad()
self.tag = utils.Scratchpad()
if not isinstance(inputs, (list, tuple)):
raise TypeError("The inputs of an Apply must be a list or tuple")
......@@ -383,7 +383,8 @@ class Variable(Node):
def __init__(self, type, owner=None, index=None, name=None):
super(Variable, self).__init__()
self.tag = utils.scratchpad()
self.tag = utils.ValidatingScratchpad("test_value", type.filter)
self.type = type
if owner is not None and not isinstance(owner, Apply):
raise TypeError("owner must be an Apply instance", owner)
......
theano/gof/op.py
浏览文件 @ 9f7a1b69
......@@ -553,29 +553,10 @@ class PureOp(object):
elif isinstance(v, SharedVariable):
return v.get_value(borrow=True, return_internal_type=True)
elif isinstance(v, graph.Variable) and hasattr(v.tag, "test_value"):
# ensure that the test value is correct
try:
ret = v.type.filter(v.tag.test_value)
except Exception as e:
# Better error message.
detailed_err_msg = (
"For compute_test_value, one input test value does not"
" have the requested type.\n"
)
detailed_err_msg += utils.get_variable_trace_string(v)
return v.tag.test_value
detailed_err_msg += (
"\nThe error when converting the test value to that"
" variable type:"
)
# We need to only have 1 args and it should be of type
# string. Otherwise, it print the tuple and so the
# new line do not get printed.
args = (detailed_err_msg,) + tuple(str(arg) for arg in e.args)
e.args = ("\n".join(args),)
raise
return ret
detailed_err_msg = utils.get_variable_trace_string(v)
raise AttributeError("%s has no test value %s" % (v, detailed_err_msg))
def __call__(self, *inputs, **kwargs):
......@@ -1057,48 +1038,13 @@ def missing_test_message(msg):
assert action in ["ignore", "off"]
def debug_error_message(msg):
"""
Displays a message saying that an error was found in some
test_values. Becomes a warning or a ValueError depending on
config.compute_test_value.
"""
action = config.compute_test_value
# this message should never be called when the debugger is off
assert action != "off"
if action in ["raise", "ignore"]:
raise ValueError(msg)
else:
assert action == "warn"
warnings.warn(msg, stacklevel=2)
def debug_assert(condition, msg=None):
"""
Customized assert with options to ignore the assert
with just a warning
"""
if msg is None:
msg = "debug_assert failed"
if not condition:
action = config.compute_test_value
if action in ["raise", "ignore"]:
raise AssertionError(msg)
else:
assert action == "warn"
warnings.warn(msg, stacklevel=2)
def get_debug_values(*args):
"""
Intended use:
for val_1, ..., val_n in get_debug_values(var_1, ..., var_n):
if some condition on val_1, ..., val_n is not met:
debug_error_message("condition was not met")
missing_test_message("condition was not met")
Given a list of variables, get_debug_values does one of three things:
......@@ -1128,10 +1074,10 @@ def get_debug_values(*args):
except AttributeError:
if hasattr(arg, "name") and arg.name is not None:
missing_test_message(
"Argument " + str(i) + "('" + arg.name + "') has no test value"
"Argument {} ('{}') has no test value".format(i, arg.name)
)
else:
missing_test_message("Argument " + str(i) + " has no test value")
missing_test_message("Argument {} has no test value".format(i))
return []
if len(rval) == 1:
......
theano/gof/utils.py
浏览文件 @ 9f7a1b69
......@@ -239,7 +239,7 @@ class object2(with_metaclass(MetaObject, object)):
return not self == other
class scratchpad(object):
class Scratchpad(object):
def clear(self):
self.__dict__.clear()
......@@ -259,6 +259,23 @@ class scratchpad(object):
print(" %s: %s" % (k, v))
class ValidatingScratchpad(Scratchpad):
"""This `Scratchpad` validates attribute values."""
def __init__(self, attr, attr_filter):
super().__init__()
object.__setattr__(self, "attr", attr)
object.__setattr__(self, "attr_filter", attr_filter)
def __setattr__(self, attr, obj):
if getattr(self, "attr", None) == attr:
obj = self.attr_filter(obj)
return object.__setattr__(self, attr, obj)
class D:
def __init__(self, **d):
self.__dict__.update(d)
......
theano/gof/vm.py
浏览文件 @ 9f7a1b69
......@@ -924,7 +924,7 @@ class VM_Linker(link.LocalLinker):
if self.use_cloop and (
self.callback is not None or self.callback_input is not None
):
logger.warn("CVM does not support callback, using Stack VM.")
logger.warning("CVM does not support callback, using Stack VM.")
if self.use_cloop and config.profile_memory:
warnings.warn("CVM does not support memory profile, using Stack VM.")
if not self.use_cloop and self.allow_partial_eval:
......
theano/misc/pkl_utils.py
浏览文件 @ 9f7a1b69
......@@ -12,6 +12,9 @@ import sys
import tempfile
import zipfile
import warnings
import theano
from collections import defaultdict
from contextlib import closing
from pickle import HIGHEST_PROTOCOL
......@@ -22,10 +25,7 @@ try:
except ImportError:
DEFAULT_PROTOCOL = HIGHEST_PROTOCOL
import theano
from theano import config
from theano.compat import PY3
from six import string_types
from theano.compile.sharedvalue import SharedVariable
__docformat__ = "restructuredtext en"
......@@ -68,7 +68,7 @@ class StripPickler(Pickler):
def save(self, obj):
# Remove the tag.trace attribute from Variable and Apply nodes
if isinstance(obj, theano.gof.utils.scratchpad):
if isinstance(obj, theano.gof.utils.Scratchpad):
for tag in self.tag_to_remove:
if hasattr(obj, tag):
del obj.__dict__[tag]
......@@ -80,93 +80,6 @@ class StripPickler(Pickler):
return Pickler.save(self, obj)
# Make an unpickler that tries encoding byte streams before raising TypeError.
# This is useful with python 3, in order to unpickle files created with
# python 2.
# This code is taken from Pandas, https://github.com/pydata/pandas,
# under the same 3-clause BSD license.
def load_reduce(self):
stack = self.stack
args = stack.pop()
func = stack[-1]
try:
value = func(*args)
except Exception:
# try to reencode the arguments
if self.encoding is not None:
new_args = []
for arg in args:
if isinstance(arg, string_types):
new_args.append(arg.encode(self.encoding))
else:
new_args.append(arg)
args = tuple(new_args)
try:
stack[-1] = func(*args)
return
except Exception:
pass
# if self.is_verbose:
# print(sys.exc_info())
# print(func, args)
raise
stack[-1] = value
if PY3:
class CompatUnpickler(pickle._Unpickler):
"""
Allow to reload in python 3 some pickled numpy ndarray.
.. versionadded:: 0.8
Examples
--------
::
with open(fname, 'rb') as fp:
if PY3:
u = CompatUnpickler(fp, encoding="latin1")
else:
u = CompatUnpickler(fp)
mat = u.load()
"""
pass
# Register `load_reduce` defined above in CompatUnpickler
CompatUnpickler.dispatch[pickle.REDUCE[0]] = load_reduce
else:
class CompatUnpickler(pickle.Unpickler):
"""
Allow to reload in python 3 some pickled numpy ndarray.
.. versionadded:: 0.8
Examples
--------
::
with open(fname, 'rb') as fp:
if PY3:
u = CompatUnpickler(fp, encoding="latin1")
else:
u = CompatUnpickler(fp)
mat = u.load()
"""
pass
class PersistentNdarrayID(object):
"""Persist ndarrays in an object by saving them to a zip file.
......
theano/printing.py
浏览文件 @ 9f7a1b69
......@@ -371,11 +371,11 @@ class Print(Op):
return (1,)
class PrinterState(gof.utils.scratchpad):
class PrinterState(gof.utils.Scratchpad):
def __init__(self, props=None, **more_props):
if props is None:
props = {}
elif isinstance(props, gof.utils.scratchpad):
elif isinstance(props, gof.utils.Scratchpad):
self.__update__(props)
else:
self.__dict__.update(props)
......@@ -862,7 +862,7 @@ def pydotprint(
):
cond = node
if cond is None:
_logger.warn(
_logger.warning(
"pydotprint: cond_highlight is set but there is no"
" IfElse node in the graph"
)
......
theano/tensor/nnet/conv.py
浏览文件 @ 9f7a1b69
......@@ -559,7 +559,7 @@ class ConvOp(OpenMPOp):
" bsize(%i). We revert it to %i. This"
" won't change the result, but may make it slower."
)
_logger.warn(warnstr, self.unroll_batch, self.bsize, new)
_logger.warning(warnstr, self.unroll_batch, self.bsize, new)
self.unroll_batch = new
......@@ -585,7 +585,7 @@ class ConvOp(OpenMPOp):
" nkern(%i). We revert it to %i. This"
" won't change the result, but may make it slower."
)
_logger.warn(warnstr, self.unroll_kern, self.nkern, new)
_logger.warning(warnstr, self.unroll_kern, self.nkern, new)
self.unroll_kern = new
self.outshp = get_conv_output_shape(
......
theano/tensor/opt.py
浏览文件 @ 9f7a1b69
......@@ -3251,7 +3251,7 @@ def merge_two_slices(slice1, len1, slice2, len2):
# sl.stop backwards
n_val = sl1.stop - 1 - sl2 * sl1.step
if config.warn.subtensor_merge_bug:
warnings.warn(
warnings.warning(
(
"Your current code is fine, but Theano versions "
"prior to 0.5rc2 might have given an incorrect result. "
......@@ -3843,7 +3843,7 @@ def local_adv_sub1_adv_inc_sub1(node):
if not inp.owner.op.set_instead_of_inc:
if config.warn.inc_subtensor1_opt:
warnings.warn(
warnings.warning(
"Your current code is fine, but Theano versions "
"between 0.7rc1 and 0.10 (or development versions "
"between Nov. 2014 and May 2017) "
......@@ -5851,7 +5851,7 @@ def local_sum_prod_div_dimshuffle(node):
break
if compatible_dims:
_logger.warn(
_logger.warning(
"WARNING: Your current code is fine, but"
" Theano versions between "
"rev. 3bd9b789f5e8 (2010-06-16) and"
......@@ -5906,7 +5906,7 @@ def local_sum_prod_div_dimshuffle(node):
if config.warn.sum_div_dimshuffle_bug and isinstance(
node.op, T.Sum
):
_logger.warn(
_logger.warning(
"WARNING: Your current code is fine,"
" but Theano versions between "
"rev. 3bd9b789f5e8 (2010-06-16) and"
......@@ -6016,7 +6016,7 @@ def local_op_of_op(node):
and newaxis != newaxis_old
and len(newaxis) == len(newaxis_old)
):
_logger.warn(
_logger.warning(
"WARNING (YOUR CURRENT CODE IS FINE): Theano "
"versions between version 9923a40c7b7a and August "
"2nd, 2010 generated bugged code in this case. "
......@@ -6102,7 +6102,7 @@ def local_reduce_join(node):
# I put this warning late to don't add extra warning.
if len(reduce_axis) != 1 or 0 not in reduce_axis:
if theano.config.warn.reduce_join:
warnings.warn(
warnings.warning(
(
"Your current code is fine, but Theano versions "
"prior to 0.7 (or this development version Sept 2014) "
......@@ -7229,19 +7229,21 @@ register_stabilize(local_erf_neg_minus_one2)
register_specialize(local_erf_neg_minus_one2)
# Stability optimization
# log(erfc(x)) => when x>threashold,
# -x**2-log(x)-.5*log(pi)+log(1-1/(2*x**2)+3/(4*x**4)-15/(8*x**6))
# for float64: threshold=26.641747557 was choosed with:
# [(i,numpy.log(scipy.special.erfc(numpy.asarray([i],dtype='float64'))))
# for i in numpy.arange(26.641747557,26.6417475571,.00000000001)]
# for float32: threshold=10.0541949, [(i,numpy.log(scipy.special.erfc(
# numpy.asarray([i],dtype='float32')))) for i in numpy.arange(
# 10.0541948,10.0541951,.0000001)]
@register_stabilize
@register_specialize
@gof.local_optimizer([T.log])
def local_log_erfc(node):
"""Stability optimization for `log(erfc(x))`.
log(erfc(x)) => when x>threshold,
-x**2-log(x)-.5*log(pi)+log(1-1/(2*x**2)+3/(4*x**4)-15/(8*x**6))
for float64: threshold=26.641747557 was choosed with:
[(i,numpy.log(scipy.special.erfc(numpy.asarray([i],dtype='float64'))))
for i in numpy.arange(26.641747557,26.6417475571,.00000000001)]
for float32: threshold=10.0541949, [(i,numpy.log(scipy.special.erfc(
numpy.asarray([i],dtype='float32')))) for i in numpy.arange(
10.0541948,10.0541951,.0000001)]
"""
if node.op != T.log:
return False
if not node.inputs[0].owner or node.inputs[0].owner.op != T.erfc:
......@@ -7270,21 +7272,26 @@ def local_log_erfc(node):
return [ret]
# Stability optimization of the grad of log(erfc(x))
# ([y*]exp(-(x**2)))/erfc(x) # The y* is optional
# ([y*]exp(x**2))/erfc(-x) => [y*](when x>threashold,
# sqrt(pi)*-x/(1-1/(2*x**2)+3/(4*x**4)-15/(8*x**6)))
# for float64: threshold=26.63 see at the end of the fct for the explanation
# for float32: threshold=9.3 see at the end of the fct for the explanation
# TODO: remove the contraint that there are only 2 inputs to exp(x**2)
# is the second.
# TODO: at the test point 10 in float32, there is instability in the original
# value. The original gives -30.0, the stab -20.1 and in float64 -18.1.
# Make it so that the test does not generate an error in that case!
@register_stabilize
@register_specialize
@gof.local_optimizer([T.true_div])
def local_grad_log_erfc_neg(node):
"""Stability optimization for the grad of `log(erfc(x))`.
([y*]exp(-(x**2)))/erfc(x) # The y* is optional
([y*]exp(x**2))/erfc(-x) => [y*](when x>threashold,
sqrt(pi)*-x/(1-1/(2*x**2)+3/(4*x**4)-15/(8*x**6)))
for float64: threshold=26.63 see at the end of the fct for the explanation
for float32: threshold=9.3 see at the end of the fct for the explanation
TODO: remove the contraint that there are only 2 inputs to exp(x**2)
is the second.
TODO: at the test point 10 in float32, there is instability in the original
value. The original gives -30.0, the stab -20.1 and in float64 -18.1.
Make it so that the test does not generate an error in that case!
"""
if node.op != T.true_div:
return False
if not node.inputs[1].owner or node.inputs[1].owner.op != T.erfc:
......@@ -7602,43 +7609,52 @@ for i in range(1,len(p64)): print i, 64[i]-p64[i-1]
"""
# ###############
# # Loop fusion #
# ###############
def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
"""
We parametrize it to make it work for Elemwise and GpuElemwise op.
def local_elemwise_fusion_op(op_class, max_input_fct=lambda node: 32, maker=None):
"""Create a recursive function that fuses `Elemwise` `Op`s.
The basic idea is that we loop through an `Elemwise` node's inputs, find
other `Elemwise` nodes, determine the scalars input types for all of the
`Elemwise` `Op`s, construct a new scalar `Op` using the scalar input types
and each `Elemwise`'s scalar `Op`, and use the composite scalar `Op` in a
new "fused" `Elemwise`.
It's parameterized in order to work for `Elemwise` and `GpuElemwise` `Op`s.
Parameters
----------
OP
GpuElemwise or Elemwise class (the one that we want to fuse)
max_input_fct
A function that returns the maximum number of inputs
that this elemwise can take (useful for GpuElemwise).
GPU kernel currently has a limit of 256 bytes for
the size of all parameters passed to it. As currently
we pass many information only by parameter, we must
limit how many ops we fuse together to avoid busting
that 256 limit.
op_class : type
`GpuElemwise` or `Elemwise` class (the one that we want to fuse)
max_input_fct : callable
A function that returns the maximum number of inputs that this `Elemwise`
can take (useful for `GpuElemwise`). The GPU kernel currently has a
limit of 256 bytes for the size of all parameters passed to it. As
currently we pass a lot of information only by parameter, we must limit how
many `Op`s we fuse together to avoid busting that 256 limit.
On the CPU we limit to 32 input variables
since that is the maximum numpy support.
On the CPU we limit to 32 input variables since that is the maximum
NumPy support.
maker: callable
A function with the signature `(node, *args)` that constructs an
`op_class` instance (e.g. `op_class(*args)`).
"""
if maker is None:
def maker(node, scalar_op):
return OP(scalar_op)
return op_class(scalar_op)
def local_fuse(node):
"""
As part of specialization, we fuse two consecutive elemwise Ops of the
"""Fuse `Elemwise` `Op`s in a node.
As part of specialization, we fuse two consecutive elemwise `Op`s of the
same shape.
For mixed dtype, we let the Composite op do the cast. It lets the C
For mixed dtype, we let the `Composite` `Op` do the cast. It lets the C
compiler do the cast.
The number of dimensions is validated at call time by theano itself.
The number of dimensions is validated at call time by Theano itself.
"""
# META TODO: PUT THESE THINGS IN TRAC, NOT TODO NOTES!!
......@@ -7665,12 +7681,13 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
# worthwhile if the summation axis doesn't line up with a
# contiguous dimension)
if type(node.op) is not OP:
if type(node.op) is not op_class:
return False
if len(node.outputs) > 1:
# We don't support the fusion for node with multiple outputs.
# We don't support fusion for nodes with multiple outputs.
return
inputs = [] # inputs of the new Elemwise op.
s_inputs = [] # inputs of the new scalar op used by the Composite.
# Inputs of the new scalar op that represents the current node.
......@@ -7691,7 +7708,6 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
for i in node.inputs:
do_fusion = False
catch = False
# Will store inputs of the fused node that are not currently inputs
# of the node we want to create (to avoid duplicating inputs).
tmp_input = []
......@@ -7704,7 +7720,7 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
# we still want to fusion. So we take the set.
if (
i.owner
and isinstance(i.owner.op, OP)
and isinstance(i.owner.op, op_class)
and len(set([n for n, idx in i.clients])) == 1
and
# Do not merge elemwise that don't have the same
......@@ -7712,7 +7728,6 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
# computation due to broadcast.
i.owner.outputs[0].broadcastable == node.outputs[0].broadcastable
):
do_fusion = True
try:
tmp_s_input = []
# we should not put duplicate input into s_inputs and inputs
......@@ -7728,12 +7743,17 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
if tv.size > 0:
tmp.tag.test_value = tv.flatten()[0]
else:
tmp.tag.test_value = tv
_logger.warning(
"Cannot construct a scalar test value"
" from a test value with no size: {}".format(ii)
)
except AttributeError:
pass
tmp_s_input.append(tmp)
tmp_input.append(ii)
tmp_scalar.append(tmp_s_input[-1])
s_op = i.owner.op.scalar_op(*tmp_s_input, return_list=True)
# if the scalar_op don't have a c implementation,
......@@ -7746,12 +7766,11 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
["z" for z in i.owner.outputs],
{"fail": "%(fail)s"},
)
except MethodNotDefined:
catch = True
except NotImplementedError:
catch = True
if catch:
_logger.info(
do_fusion = True
except (NotImplementedError, MethodNotDefined):
_logger.warning(
(
"%s does not implement the c_code function."
" As well as being potentially slow, this"
......@@ -7782,8 +7801,8 @@ def local_elemwise_fusion_op(OP, max_input_fct=lambda node: 32, maker=None):
s_inputs.extend(tmp_scalar)
s_g.extend(s_op)
else:
# We must support the case where the same variable appear many
# time in the inputs
# We must support the case where the same variable appears many
# times within the inputs
if inputs.count(i) == node.inputs.count(i):
s = s_inputs[inputs.index(i)]
else:
......@@ -7819,8 +7838,8 @@ your code will run correctly, but may be slower."""
["z" for x in s_new_out],
{"fail": "%(fail)s"},
)
except MethodNotDefined:
_logger.info(
except (NotImplementedError, MethodNotDefined):
_logger.warning(
(
"%s does not implement the c_code function."
" As well as being potentially slow, this disables "
......@@ -7828,29 +7847,19 @@ your code will run correctly, but may be slower."""
)
% str(s_new_out[0].owner.op)
)
return False
except NotImplementedError:
_logger.info(
(
"%s does not implement the c_code function. As well"
" as being potentially slow, this disables loop"
" fusion of this op."
)
% str(s_new_out[0].owner.op)
)
return False
# create the composite op.
C = scalar.Composite(s_inputs, s_new_out)
composite_op = scalar.Composite(s_inputs, s_new_out)
# create the new node.
# Do not call make_node to have test_value
n = maker(node, C)(*inputs).owner
assert len(n.outputs) == 1
assert node.outputs[0].dtype == n.outputs[0].dtype
new_node = maker(node, composite_op)(*inputs).owner
assert len(new_node.outputs) == 1
assert node.outputs[0].dtype == new_node.outputs[0].dtype
if len(n.inputs) > max_nb_input:
_logger.info(
if len(new_node.inputs) > max_nb_input:
_logger.warning(
"loop fusion failed because Op would exceed" " kernel argument limit."
)
return False
......@@ -7858,16 +7867,15 @@ your code will run correctly, but may be slower."""
# we fuse as many that we can at the same time to make debug mode faster
# debug mode will be faster as it won't test all intermediate step.
while True:
ret = local_fuse(n)
ret = local_fuse(new_node)
if ret is not False and ret is not None:
# print n,ret
assert len(ret) == len(n.outputs)
assert len(ret) == len(new_node.outputs)
assert len(ret) == 1
n = ret[0].owner
new_node = ret[0].owner
else:
break
return n.outputs
return new_node.outputs
return local_fuse
......
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