Skip to content

P
pytensor
提交 13cb7ed2 authored 作者: amrithasuresh's avatar amrithasuresh
浏览文件
操作

Updated numpy as np

上级 c918af6e
显示空白字符变更
内嵌 并排
正在显示 包含 123 行增加 和 123 行删除
theano/tensor/nnet/tests/test_nnet.py
浏览文件 @ 13cb7ed2
from __future__ import absolute_import, print_function, division from __future__ import absolute_import, print_function, division
import unittest import unittest
import numpy import numpy as np
from nose.plugins.skip import SkipTest from nose.plugins.skip import SkipTest
from six.moves import xrange from six.moves import xrange
...@@ -47,7 +47,7 @@ class T_sigmoid(unittest.TestCase): ...@@ -47,7 +47,7 @@ class T_sigmoid(unittest.TestCase):
utt.seed_rng() utt.seed_rng()
def test_elemwise(self): def test_elemwise(self):
utt.verify_grad(sigmoid, [numpy.random.rand(3, 4)]) utt.verify_grad(sigmoid, [np.random.rand(3, 4)])
class T_softplus(unittest.TestCase): class T_softplus(unittest.TestCase):
...@@ -56,7 +56,7 @@ class T_softplus(unittest.TestCase): ...@@ -56,7 +56,7 @@ class T_softplus(unittest.TestCase):
utt.seed_rng() utt.seed_rng()
def test_elemwise(self): def test_elemwise(self):
utt.verify_grad(softplus, [numpy.random.rand(3, 4)]) utt.verify_grad(softplus, [np.random.rand(3, 4)])
class T_Softmax(utt.InferShapeTester): class T_Softmax(utt.InferShapeTester):
...@@ -64,26 +64,26 @@ class T_Softmax(utt.InferShapeTester): ...@@ -64,26 +64,26 @@ class T_Softmax(utt.InferShapeTester):
def test0(self): def test0(self):
def f(a): def f(a):
return softmax_op(a)[:, 0] return softmax_op(a)[:, 0]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test1(self): def test1(self):
def f(a): def f(a):
return softmax_op(a)[:, 1] return softmax_op(a)[:, 1]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test2(self): def test2(self):
def f(a): def f(a):
return softmax_op(a)[:, 2] return softmax_op(a)[:, 2]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test3(self): def test3(self):
def f(a): def f(a):
return softmax_op(a)[:, 3] return softmax_op(a)[:, 3]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
admat_val = numpy.random.rand(3, 4).astype(config.floatX) admat_val = np.random.rand(3, 4).astype(config.floatX)
self._compile_and_check([admat], [Softmax()(admat)], self._compile_and_check([admat], [Softmax()(admat)],
[admat_val], Softmax) [admat_val], Softmax)
...@@ -91,13 +91,13 @@ class T_Softmax(utt.InferShapeTester): ...@@ -91,13 +91,13 @@ class T_Softmax(utt.InferShapeTester):
x = T.vector() x = T.vector()
f = theano.function([x], softmax_op(x)) f = theano.function([x], softmax_op(x))
xv = numpy.random.randn(6).astype(config.floatX) xv = np.random.randn(6).astype(config.floatX)
assert numpy.allclose(f(xv), numpy.exp(xv) / numpy.exp(xv).sum()) assert np.allclose(f(xv), np.exp(xv) / np.exp(xv).sum())
def test_vector_grad(self): def test_vector_grad(self):
def f(a): def f(a):
return softmax_op(a) return softmax_op(a)
utt.verify_grad(f, [numpy.random.rand(4)]) utt.verify_grad(f, [np.random.rand(4)])
class T_SoftmaxWithBias(utt.InferShapeTester): class T_SoftmaxWithBias(utt.InferShapeTester):
...@@ -105,32 +105,32 @@ class T_SoftmaxWithBias(utt.InferShapeTester): ...@@ -105,32 +105,32 @@ class T_SoftmaxWithBias(utt.InferShapeTester):
def test0(self): def test0(self):
def f(a, b): def f(a, b):
return softmax_with_bias(a, b)[:, 0] return softmax_with_bias(a, b)[:, 0]
utt.verify_grad(f, [numpy.random.rand(3, 4), utt.verify_grad(f, [np.random.rand(3, 4),
numpy.random.rand(4)]) np.random.rand(4)])
def test1(self): def test1(self):
def f(a, b): def f(a, b):
return softmax_with_bias(a, b)[:, 1] return softmax_with_bias(a, b)[:, 1]
utt.verify_grad(f, [numpy.random.rand(3, 4), utt.verify_grad(f, [np.random.rand(3, 4),
numpy.random.rand(4)]) np.random.rand(4)])
def test2(self): def test2(self):
def f(a, b): def f(a, b):
return softmax_with_bias(a, b)[:, 2] return softmax_with_bias(a, b)[:, 2]
utt.verify_grad(f, [numpy.random.rand(3, 4), utt.verify_grad(f, [np.random.rand(3, 4),
numpy.random.rand(4)]) np.random.rand(4)])
def test3(self): def test3(self):
def f(a, b): def f(a, b):
return softmax_with_bias(a, b)[:, 3] return softmax_with_bias(a, b)[:, 3]
utt.verify_grad(f, [numpy.random.rand(3, 4), utt.verify_grad(f, [np.random.rand(3, 4),
numpy.random.rand(4)]) np.random.rand(4)])
def test_broadcast(self): def test_broadcast(self):
# test that we don't raise an error during optimization for no good # test that we don't raise an error during optimization for no good
# reason as softmax_with_bias don't support correctly some/all # reason as softmax_with_bias don't support correctly some/all
# broadcasted inputs pattern # broadcasted inputs pattern
initial_W = numpy.asarray([[0.1, 0.1, 0.1], initial_W = np.asarray([[0.1, 0.1, 0.1],
[0.1, 0.1, 0.1], [0.1, 0.1, 0.1],
[0.1, 0.1, 0.1]], [0.1, 0.1, 0.1]],
dtype=theano.config.floatX) dtype=theano.config.floatX)
...@@ -148,8 +148,8 @@ class T_SoftmaxWithBias(utt.InferShapeTester): ...@@ -148,8 +148,8 @@ class T_SoftmaxWithBias(utt.InferShapeTester):
def test_softmax_with_bias_trace(self): def test_softmax_with_bias_trace(self):
a = theano.shared( a = theano.shared(
numpy.random.randn(3).astype(config.floatX)) np.random.randn(3).astype(config.floatX))
b = theano.shared(numpy.float32(numpy.random.randn())) b = theano.shared(np.float32(np.random.randn()))
sm = T.nnet.softmax(a + b) sm = T.nnet.softmax(a + b)
f = theano.function([], sm) f = theano.function([], sm)
assert check_stack_trace(f, ops_to_check='last') assert check_stack_trace(f, ops_to_check='last')
...@@ -157,8 +157,8 @@ class T_SoftmaxWithBias(utt.InferShapeTester): ...@@ -157,8 +157,8 @@ class T_SoftmaxWithBias(utt.InferShapeTester):
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
advec = vector() advec = vector()
admat_val = numpy.random.rand(3, 4).astype(config.floatX) admat_val = np.random.rand(3, 4).astype(config.floatX)
advec_val = numpy.random.rand(4).astype(config.floatX) advec_val = np.random.rand(4).astype(config.floatX)
self._compile_and_check([admat, advec], self._compile_and_check([admat, advec],
[SoftmaxWithBias()(admat, advec)], [SoftmaxWithBias()(admat, advec)],
[admat_val, advec_val], SoftmaxWithBias) [admat_val, advec_val], SoftmaxWithBias)
...@@ -169,40 +169,40 @@ class T_LogSoftmax(utt.InferShapeTester): ...@@ -169,40 +169,40 @@ class T_LogSoftmax(utt.InferShapeTester):
def test0(self): def test0(self):
def f(a): def f(a):
return logsoftmax_op(a)[:, 0] return logsoftmax_op(a)[:, 0]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test1(self): def test1(self):
def f(a): def f(a):
return logsoftmax_op(a)[:, 1] return logsoftmax_op(a)[:, 1]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test2(self): def test2(self):
def f(a): def f(a):
return logsoftmax_op(a)[:, 2] return logsoftmax_op(a)[:, 2]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test3(self): def test3(self):
def f(a): def f(a):
return logsoftmax_op(a)[:, 3] return logsoftmax_op(a)[:, 3]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test_matrix(self): def test_matrix(self):
def f(a): def f(a):
return logsoftmax_op(a) return logsoftmax_op(a)
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test_vector(self): def test_vector(self):
x = T.vector() x = T.vector()
f = theano.function([x], logsoftmax_op(x)) f = theano.function([x], logsoftmax_op(x))
xv = numpy.random.randn(6).astype(config.floatX) xv = np.random.randn(6).astype(config.floatX)
assert numpy.allclose(f(xv), assert np.allclose(f(xv),
numpy.log(numpy.exp(xv) / numpy.exp(xv).sum())) np.log(np.exp(xv) / np.exp(xv).sum()))
def test_vector_grad(self): def test_vector_grad(self):
def f(a): def f(a):
return logsoftmax_op(a) return logsoftmax_op(a)
utt.verify_grad(f, [numpy.random.rand(4)]) utt.verify_grad(f, [np.random.rand(4)])
def test_allclose(self): def test_allclose(self):
m = theano.config.mode m = theano.config.mode
...@@ -220,9 +220,9 @@ class T_LogSoftmax(utt.InferShapeTester): ...@@ -220,9 +220,9 @@ class T_LogSoftmax(utt.InferShapeTester):
grad = tensor.grad(cm2.mean(), x) grad = tensor.grad(cm2.mean(), x)
# create some inputs into a softmax that are large and labels # create some inputs into a softmax that are large and labels
a = numpy.exp(10 * numpy.random.rand(5, 10).astype(theano.config.floatX)) a = np.exp(10 * np.random.rand(5, 10).astype(theano.config.floatX))
# create some one-hot coded labels # create some one-hot coded labels
b = numpy.eye(5, 10).astype(theano.config.floatX) b = np.eye(5, 10).astype(theano.config.floatX)
# show equivalence of softmax and exponentiated numerically stable # show equivalence of softmax and exponentiated numerically stable
# log-softmax # log-softmax
...@@ -241,7 +241,7 @@ class T_LogSoftmax(utt.InferShapeTester): ...@@ -241,7 +241,7 @@ class T_LogSoftmax(utt.InferShapeTester):
# while in the log-softmax case they don't # while in the log-softmax case they don't
f3 = theano.function([x, y], [grad]) f3 = theano.function([x, y], [grad])
grad_ = f3(a, b) grad_ = f3(a, b)
assert not numpy.any(numpy.isnan(grad_)) assert not np.any(np.isnan(grad_))
def test_isclose(self): def test_isclose(self):
def f(a): def f(a):
...@@ -274,8 +274,8 @@ class T_LogSoftmax(utt.InferShapeTester): ...@@ -274,8 +274,8 @@ class T_LogSoftmax(utt.InferShapeTester):
m.check_isfinite = False m.check_isfinite = False
# some inputs that are large to make the gradient explode in the non # some inputs that are large to make the gradient explode in the non
# optimized case # optimized case
a = numpy.exp( a = np.exp(
10 * numpy.random.rand(5, 10).astype(theano.config.floatX)) 10 * np.random.rand(5, 10).astype(theano.config.floatX))
def myfunc(x): def myfunc(x):
sm = tensor.nnet.softmax(x) sm = tensor.nnet.softmax(x)
...@@ -317,8 +317,8 @@ class T_SoftmaxGrad(utt.InferShapeTester): ...@@ -317,8 +317,8 @@ class T_SoftmaxGrad(utt.InferShapeTester):
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
bdmat = matrix() bdmat = matrix()
admat_val = numpy.random.rand(3, 4).astype(config.floatX) admat_val = np.random.rand(3, 4).astype(config.floatX)
bdmat_val = numpy.random.rand(3, 4).astype(config.floatX) bdmat_val = np.random.rand(3, 4).astype(config.floatX)
self._compile_and_check([admat, bdmat], [SoftmaxGrad()(admat, bdmat)], self._compile_and_check([admat, bdmat], [SoftmaxGrad()(admat, bdmat)],
[admat_val, bdmat_val], SoftmaxGrad) [admat_val, bdmat_val], SoftmaxGrad)
...@@ -333,29 +333,29 @@ class T_CrossentropySoftmax1Hot(unittest.TestCase): ...@@ -333,29 +333,29 @@ class T_CrossentropySoftmax1Hot(unittest.TestCase):
def f(a, b): def f(a, b):
return crossentropy_softmax_1hot_with_bias(a, b, y_idx)[0] return crossentropy_softmax_1hot_with_bias(a, b, y_idx)[0]
utt.verify_grad(f, [numpy.random.rand(3, 4), utt.verify_grad(f, [np.random.rand(3, 4),
numpy.random.rand(4)]) np.random.rand(4)])
def test1(self): def test1(self):
y_idx = [0, 1, 3] y_idx = [0, 1, 3]
def f(a): def f(a):
return crossentropy_softmax_1hot(a, y_idx)[0] return crossentropy_softmax_1hot(a, y_idx)[0]
utt.verify_grad(f, [numpy.random.rand(3, 4)]) utt.verify_grad(f, [np.random.rand(3, 4)])
def test_vector(self): def test_vector(self):
y_idx = [3] y_idx = [3]
def f(a): def f(a):
return crossentropy_softmax_1hot(T.shape_padleft(a), y_idx)[0] return crossentropy_softmax_1hot(T.shape_padleft(a), y_idx)[0]
utt.verify_grad(f, [numpy.random.rand(4)]) utt.verify_grad(f, [np.random.rand(4)])
def test_vectors(self): def test_vectors(self):
y_idx = [3] y_idx = [3]
def f(a, b): def f(a, b):
return crossentropy_softmax_1hot(T.shape_padleft(a) + b, y_idx)[0] return crossentropy_softmax_1hot(T.shape_padleft(a) + b, y_idx)[0]
utt.verify_grad(f, [numpy.random.rand(4), numpy.random.rand(4)]) utt.verify_grad(f, [np.random.rand(4), np.random.rand(4)])
class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester): class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester):
...@@ -364,20 +364,20 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester): ...@@ -364,20 +364,20 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester):
def ff(class_dtype): def ff(class_dtype):
def f(sm): def f(sm):
# Class indices # Class indices
y = numpy.random.randint(low=0, high=5, size=10).astype(class_dtype) y = np.random.randint(low=0, high=5, size=10).astype(class_dtype)
return theano.tensor.nnet.crossentropy_softmax_1hot_with_bias_dx( return theano.tensor.nnet.crossentropy_softmax_1hot_with_bias_dx(
numpy.random.rand(10), # Gradient w.r.t. NLL. np.random.rand(10), # Gradient w.r.t. NLL.
sm, # Softmax output. sm, # Softmax output.
y) y)
return f return f
# Build a random softmax output whose rows sum to 1. # Build a random softmax output whose rows sum to 1.
softmax_output = numpy.random.rand(10, 5) softmax_output = np.random.rand(10, 5)
softmax_output /= softmax_output.sum(axis=1).reshape(10, 1) softmax_output /= softmax_output.sum(axis=1).reshape(10, 1)
for dtype in ['uint8', 'int8', 'uint64', 'int64']: for dtype in ['uint8', 'int8', 'uint64', 'int64']:
utt.verify_grad(ff(dtype), [softmax_output]) utt.verify_grad(ff(dtype), [softmax_output])
def test1(self): def test1(self):
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
softmax_output = rng.rand(10, 5) softmax_output = rng.rand(10, 5)
softmax_output /= softmax_output.sum(axis=1).reshape(10, 1) softmax_output /= softmax_output.sum(axis=1).reshape(10, 1)
...@@ -392,7 +392,7 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester): ...@@ -392,7 +392,7 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester):
admat = matrix() admat = matrix()
advec = vector() advec = vector()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(10, 5).astype(config.floatX) admat_val = rng.rand(10, 5).astype(config.floatX)
admat_val /= admat_val.sum(axis=1).reshape(10, 1) admat_val /= admat_val.sum(axis=1).reshape(10, 1)
advec_val = rng.rand(10).astype(config.floatX) advec_val = rng.rand(10).astype(config.floatX)
...@@ -407,7 +407,7 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester): ...@@ -407,7 +407,7 @@ class T_CrossentropySoftmax1HotWithBiasDx(utt.InferShapeTester):
admat = matrix() admat = matrix()
advec = vector() advec = vector()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(10, 5).astype(config.floatX) admat_val = rng.rand(10, 5).astype(config.floatX)
admat_val /= admat_val.sum(axis=1).reshape(10, 1) admat_val /= admat_val.sum(axis=1).reshape(10, 1)
advec_val = rng.rand(10).astype(config.floatX) advec_val = rng.rand(10).astype(config.floatX)
...@@ -431,28 +431,28 @@ class T_CrossentropySoftmaxArgmax1HotWithBias(utt.InferShapeTester): ...@@ -431,28 +431,28 @@ class T_CrossentropySoftmaxArgmax1HotWithBias(utt.InferShapeTester):
# First test gradient when getting a gradient on the NLL output. # First test gradient when getting a gradient on the NLL output.
def grad_on_nll_dtype(dtype): def grad_on_nll_dtype(dtype):
def grad_on_nll(x, b): def grad_on_nll(x, b):
y_idx = numpy.random.randint(low=0, high=n_classes, size=n_samples).astype(dtype) y_idx = np.random.randint(low=0, high=n_classes, size=n_samples).astype(dtype)
return self.op(x, b, y_idx=y_idx)[0] return self.op(x, b, y_idx=y_idx)[0]
return grad_on_nll return grad_on_nll
for dtype in ['uint8', 'int8', 'uint64', 'int64']: for dtype in ['uint8', 'int8', 'uint64', 'int64']:
utt.verify_grad(grad_on_nll_dtype(dtype), utt.verify_grad(grad_on_nll_dtype(dtype),
[numpy.random.rand(n_samples, n_classes), [np.random.rand(n_samples, n_classes),
numpy.random.rand(n_classes)]) np.random.rand(n_classes)])
# Then test gradient when getting a gradient on the softmax output. # Then test gradient when getting a gradient on the softmax output.
def grad_on_softmax(x, b): def grad_on_softmax(x, b):
return self.op(x, b, y_idx=numpy.random.randint( return self.op(x, b, y_idx=np.random.randint(
low=0, high=n_classes, size=n_samples))[1] low=0, high=n_classes, size=n_samples))[1]
utt.verify_grad( utt.verify_grad(
grad_on_softmax, grad_on_softmax,
[numpy.random.rand(n_samples, n_classes), [np.random.rand(n_samples, n_classes),
numpy.random.rand(n_classes)]) np.random.rand(n_classes)])
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
advec = vector() advec = vector()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(3, 5).astype(config.floatX) admat_val = rng.rand(3, 5).astype(config.floatX)
advec_val = rng.rand(5).astype(config.floatX) advec_val = rng.rand(5).astype(config.floatX)
alvec_val = rng.randint(low=0, high=5, size=3) alvec_val = rng.randint(low=0, high=5, size=3)
...@@ -466,7 +466,7 @@ class T_CrossentropySoftmaxArgmax1HotWithBias(utt.InferShapeTester): ...@@ -466,7 +466,7 @@ class T_CrossentropySoftmaxArgmax1HotWithBias(utt.InferShapeTester):
admat = matrix() admat = matrix()
advec = vector() advec = vector()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(3, 5).astype(config.floatX) admat_val = rng.rand(3, 5).astype(config.floatX)
advec_val = rng.rand(5).astype(config.floatX) advec_val = rng.rand(5).astype(config.floatX)
alvec_val = rng.randint(low=0, high=5, size=3) alvec_val = rng.randint(low=0, high=5, size=3)
...@@ -482,27 +482,27 @@ class T_prepend(utt.InferShapeTester): ...@@ -482,27 +482,27 @@ class T_prepend(utt.InferShapeTester):
x = tensor.matrix('x') x = tensor.matrix('x')
y = Prepend_scalar_constant_to_each_row(4.)(x) y = Prepend_scalar_constant_to_each_row(4.)(x)
f = theano.function([x], y) f = theano.function([x], y)
m = numpy.random.rand(3, 5).astype(config.floatX) m = np.random.rand(3, 5).astype(config.floatX)
my = f(m) my = f(m)
self.assertTrue(my.shape == (3, 6), my.shape) self.assertTrue(my.shape == (3, 6), my.shape)
self.assertTrue(numpy.all(my[:, 0] == 4.0)) self.assertTrue(np.all(my[:, 0] == 4.0))
def test1(self): def test1(self):
"basic functionality" "basic functionality"
x = tensor.matrix('x') x = tensor.matrix('x')
y = Prepend_scalar_to_each_row()(5., x) y = Prepend_scalar_to_each_row()(5., x)
f = theano.function([x], y) f = theano.function([x], y)
m = numpy.ones((3, 5), dtype="float32") m = np.ones((3, 5), dtype="float32")
my = f(m) my = f(m)
self.assertTrue(my.shape == (3, 6)) self.assertTrue(my.shape == (3, 6))
self.assertTrue(numpy.all(my[:, 0] == 5.0)) self.assertTrue(np.all(my[:, 0] == 5.0))
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
adscal = scalar() adscal = scalar()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(3, 5).astype(config.floatX) admat_val = rng.rand(3, 5).astype(config.floatX)
adscal_val = numpy.asarray(rng.rand(), dtype=config.floatX).item() adscal_val = np.asarray(rng.rand(), dtype=config.floatX).item()
self._compile_and_check( self._compile_and_check(
[admat], [admat],
[Prepend_scalar_constant_to_each_row(adscal_val)(admat)], [Prepend_scalar_constant_to_each_row(adscal_val)(admat)],
...@@ -522,7 +522,7 @@ class T_CrossentropyCategorical1HotGrad(utt.InferShapeTester): ...@@ -522,7 +522,7 @@ class T_CrossentropyCategorical1HotGrad(utt.InferShapeTester):
advec = vector() advec = vector()
admat = matrix() admat = matrix()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
advec_val = rng.rand(3).astype(config.floatX) advec_val = rng.rand(3).astype(config.floatX)
admat_val = rng.rand(3, 2).astype(config.floatX) admat_val = rng.rand(3, 2).astype(config.floatX)
alvec_val = [0, 1, 0] alvec_val = [0, 1, 0]
...@@ -541,21 +541,21 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -541,21 +541,21 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
op = crossentropy_categorical_1hot op = crossentropy_categorical_1hot
xe = op(x, one_of_n) xe = op(x, one_of_n)
f = theano.function([x, one_of_n], xe) f = theano.function([x, one_of_n], xe)
x_val = numpy.asarray( x_val = np.asarray(
[[.4, .6, .0], [.1, .8, .1]], [[.4, .6, .0], [.1, .8, .1]],
dtype=config.floatX) dtype=config.floatX)
xe_val = f(x_val, [0, 1]) xe_val = f(x_val, [0, 1])
assert numpy.allclose(xe_val, -numpy.log([.4, .8])) assert np.allclose(xe_val, -np.log([.4, .8]))
def oplike(x): def oplike(x):
return op(x, [0, 1]) return op(x, [0, 1])
tensor.verify_grad(oplike, [x_val], rng=numpy.random) tensor.verify_grad(oplike, [x_val], rng=np.random)
def test_infer_shape(self): def test_infer_shape(self):
admat = matrix() admat = matrix()
alvec = lvector() alvec = lvector()
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
admat_val = rng.rand(3, 2).astype(config.floatX) admat_val = rng.rand(3, 2).astype(config.floatX)
alvec_val = [0, 1, 0] alvec_val = [0, 1, 0]
self._compile_and_check( self._compile_and_check(
...@@ -775,10 +775,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -775,10 +775,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(3, 5).astype(config.floatX) x_val = rng.randn(3, 5).astype(config.floatX)
b_val = rng.randn(5).astype(config.floatX) b_val = rng.randn(5).astype(config.floatX)
y_val = numpy.asarray([2, 4, 1]) y_val = np.asarray([2, 4, 1])
x = T.matrix('x') x = T.matrix('x')
b = T.vector('b') b = T.vector('b')
y = T.lvector('y') y = T.lvector('y')
...@@ -954,9 +954,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -954,9 +954,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(3, 5).astype(config.floatX) x_val = rng.randn(3, 5).astype(config.floatX)
y_val = numpy.asarray([2, 4, 1], dtype='int64') y_val = np.asarray([2, 4, 1], dtype='int64')
x = T.matrix('x') x = T.matrix('x')
y = T.lvector('y') y = T.lvector('y')
yi = T.cast(y, 'int32') yi = T.cast(y, 'int32')
...@@ -1002,9 +1002,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -1002,9 +1002,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(5).astype(config.floatX) x_val = rng.randn(5).astype(config.floatX)
y_val = numpy.asarray([2]) y_val = np.asarray([2])
x = T.vector('x') x = T.vector('x')
y = T.lvector('y') y = T.lvector('y')
...@@ -1047,10 +1047,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -1047,10 +1047,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(5).astype(config.floatX) x_val = rng.randn(5).astype(config.floatX)
b_val = rng.randn(5).astype(config.floatX) b_val = rng.randn(5).astype(config.floatX)
y_val = numpy.asarray([2]) y_val = np.asarray([2])
x = T.vector('x') x = T.vector('x')
b = T.vector('b') b = T.vector('b')
...@@ -1107,10 +1107,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -1107,10 +1107,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(5).astype(config.floatX) x_val = rng.randn(5).astype(config.floatX)
b_val = rng.randn(5).astype(config.floatX) b_val = rng.randn(5).astype(config.floatX)
y_val = numpy.asarray([2]) y_val = np.asarray([2])
x = T.vector('x') x = T.vector('x')
b = T.vector('b') b = T.vector('b')
...@@ -1169,10 +1169,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -1169,10 +1169,10 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(5).astype(config.floatX) x_val = rng.randn(5).astype(config.floatX)
b_val = rng.randn(5).astype(config.floatX) b_val = rng.randn(5).astype(config.floatX)
y_val = numpy.asarray([2]) y_val = np.asarray([2])
x = T.vector('x') x = T.vector('x')
b = T.vector('b') b = T.vector('b')
...@@ -1228,9 +1228,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester): ...@@ -1228,9 +1228,9 @@ class T_CrossentropyCategorical1Hot(utt.InferShapeTester):
mode = theano.compile.mode.get_default_mode() mode = theano.compile.mode.get_default_mode()
if mode == theano.compile.mode.get_mode('FAST_COMPILE'): if mode == theano.compile.mode.get_mode('FAST_COMPILE'):
mode = 'FAST_RUN' mode = 'FAST_RUN'
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x_val = rng.randn(3, 5).astype(config.floatX) x_val = rng.randn(3, 5).astype(config.floatX)
y_val = numpy.asarray([2, 4, 1]) y_val = np.asarray([2, 4, 1])
x = T.matrix('x') x = T.matrix('x')
y = T.lvector('y') y = T.lvector('y')
a = T.scalar('a') a = T.scalar('a')
...@@ -1445,21 +1445,21 @@ def test_asymptotic_32(): ...@@ -1445,21 +1445,21 @@ def test_asymptotic_32():
for i, n in enumerate(f.maker.fgraph.toposort()): for i, n in enumerate(f.maker.fgraph.toposort()):
print(i, n) print(i, n)
xval = numpy.zeros((5, 5), dtype=dtype).astype(dtype) xval = np.zeros((5, 5), dtype=dtype).astype(dtype)
x2val = numpy.zeros(5, dtype=xval.dtype).astype(dtype) x2val = np.zeros(5, dtype=xval.dtype).astype(dtype)
for i in xrange(100): for i in xrange(100):
cval, gxval = f(xval, numpy.arange(5), x2val) cval, gxval = f(xval, np.arange(5), x2val)
xval -= 100.3 * gxval xval -= 100.3 * gxval
# print cval, gxval # print cval, gxval
assert cval == 0 # no problem going to zero error assert cval == 0 # no problem going to zero error
# what about when x gets really big? # what about when x gets really big?
xval = numpy.zeros((5, 5), dtype=dtype) xval = np.zeros((5, 5), dtype=dtype)
x2val = numpy.zeros(5, dtype=xval.dtype) x2val = np.zeros(5, dtype=xval.dtype)
for i in xrange(100): for i in xrange(100):
cval, gxval = f(xval, numpy.arange(5), x2val) cval, gxval = f(xval, np.arange(5), x2val)
xval += 100000.3 * gxval xval += 100000.3 * gxval
# print cval, gxval # print cval, gxval
...@@ -1480,7 +1480,7 @@ class Test_softmax_opt: ...@@ -1480,7 +1480,7 @@ class Test_softmax_opt:
def setUp(self): def setUp(self):
utt.seed_rng() utt.seed_rng()
self.rng = numpy.random.RandomState(utt.fetch_seed()) self.rng = np.random.RandomState(utt.fetch_seed())
self.mode = theano.compile.mode.get_default_mode() self.mode = theano.compile.mode.get_default_mode()
self.mode = self.mode.including('canonicalize') self.mode = self.mode.including('canonicalize')
...@@ -1584,7 +1584,7 @@ class Test_softmax_opt: ...@@ -1584,7 +1584,7 @@ class Test_softmax_opt:
def test_softmax_graph(): def test_softmax_graph():
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x = theano.shared(rng.normal(size=(3, 4))) x = theano.shared(rng.normal(size=(3, 4)))
def f(inputs): def f(inputs):
...@@ -1595,7 +1595,7 @@ def test_softmax_graph(): ...@@ -1595,7 +1595,7 @@ def test_softmax_graph():
def test_grad_softmax_grad(): def test_grad_softmax_grad():
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
x = theano.shared(rng.normal(size=(3, 4))) x = theano.shared(rng.normal(size=(3, 4)))
def f(inputs): def f(inputs):
...@@ -1621,39 +1621,39 @@ def test_stabilize_log_softmax(): ...@@ -1621,39 +1621,39 @@ def test_stabilize_log_softmax():
# call the function so debug mode can verify the optimized # call the function so debug mode can verify the optimized
# version matches the unoptimized version # version matches the unoptimized version
rng = numpy.random.RandomState([2012, 8, 22]) rng = np.random.RandomState([2012, 8, 22])
f(numpy.cast[config.floatX](rng.randn(2, 3))) f(np.cast[config.floatX](rng.randn(2, 3)))
def test_relu(): def test_relu():
x = matrix('x') x = matrix('x')
seed = theano.tests.unittest_tools.fetch_seed() seed = theano.tests.unittest_tools.fetch_seed()
rng = numpy.random.RandomState(seed) rng = np.random.RandomState(seed)
X = rng.randn(20, 30).astype(config.floatX) X = rng.randn(20, 30).astype(config.floatX)
# test the base case, without custom alpha value # test the base case, without custom alpha value
y = relu(x).eval({x: X}) y = relu(x).eval({x: X})
assert numpy.allclose(y, numpy.maximum(X, 0)) assert np.allclose(y, np.maximum(X, 0))
# test for different constant alpha values (also outside of [0, 1]) # test for different constant alpha values (also outside of [0, 1])
for alpha in 0, 0.3, 1, 2, -0.3, -1, -2: for alpha in 0, 0.3, 1, 2, -0.3, -1, -2:
y = relu(x, alpha).eval({x: X}) y = relu(x, alpha).eval({x: X})
assert numpy.allclose(y, numpy.where(X > 0, X, alpha * X)) assert np.allclose(y, np.where(X > 0, X, alpha * X))
# test for variable alpha (scalar, vector and matrix) # test for variable alpha (scalar, vector and matrix)
for alpha in scalar(), vector(), matrix(): for alpha in scalar(), vector(), matrix():
# create value for alpha (correct ndim and broadcastable against X) # create value for alpha (correct ndim and broadcastable against X)
A = numpy.array(rng.randn(*X.shape[::-1][:alpha.ndim][::-1]), A = np.array(rng.randn(*X.shape[::-1][:alpha.ndim][::-1]),
dtype=config.floatX) dtype=config.floatX)
y = relu(x, alpha).eval({x: X, alpha: A}) y = relu(x, alpha).eval({x: X, alpha: A})
assert numpy.allclose(y, numpy.where(X > 0, X, A * X), rtol=3e-5) assert np.allclose(y, np.where(X > 0, X, A * X), rtol=3e-5)
# test that for alpha of ndarray don't cause upcast. # test that for alpha of ndarray don't cause upcast.
x = matrix('x', dtype='float32') x = matrix('x', dtype='float32')
rng = numpy.random.RandomState(seed) rng = np.random.RandomState(seed)
X = rng.randn(20, 30).astype('float32') X = rng.randn(20, 30).astype('float32')
alpha = numpy.asarray(.123, dtype='float32') alpha = np.asarray(.123, dtype='float32')
y = relu(x, alpha).eval({x: X}) y = relu(x, alpha).eval({x: X})
assert numpy.allclose(y, numpy.where(X > 0, X, alpha * X)) assert np.allclose(y, np.where(X > 0, X, alpha * X))
assert y.dtype == 'float32' assert y.dtype == 'float32'
...@@ -1681,19 +1681,19 @@ def test_h_softmax(): ...@@ -1681,19 +1681,19 @@ def test_h_softmax():
shared = theano.shared shared = theano.shared
# First level of h_softmax # First level of h_softmax
W1 = numpy.asarray(numpy.random.normal( W1 = np.asarray(np.random.normal(
size=(input_size, h_softmax_level1_size)), dtype=floatX) size=(input_size, h_softmax_level1_size)), dtype=floatX)
W1 = shared(W1) W1 = shared(W1)
b1 = shared(numpy.asarray(numpy.zeros((h_softmax_level1_size,)), b1 = shared(np.asarray(np.zeros((h_softmax_level1_size,)),
dtype=floatX)) dtype=floatX))
# Second level of h_softmax # Second level of h_softmax
W2 = numpy.asarray(numpy.random.normal( W2 = np.asarray(np.random.normal(
size=(h_softmax_level1_size, input_size, h_softmax_level2_size)), size=(h_softmax_level1_size, input_size, h_softmax_level2_size)),
dtype=floatX) dtype=floatX)
W2 = shared(W2) W2 = shared(W2)
b2 = shared( b2 = shared(
numpy.asarray(numpy.zeros((h_softmax_level1_size, np.asarray(np.zeros((h_softmax_level1_size,
h_softmax_level2_size)), dtype=floatX)) h_softmax_level2_size)), dtype=floatX))
############# #############
...@@ -1719,8 +1719,8 @@ def test_h_softmax(): ...@@ -1719,8 +1719,8 @@ def test_h_softmax():
############# #############
# Test # Test
############# #############
x_mat = numpy.random.normal(size=(batch_size, input_size)).astype(floatX) x_mat = np.random.normal(size=(batch_size, input_size)).astype(floatX)
y_mat = numpy.random.randint(0, output_size, batch_size).astype('int32') y_mat = np.random.randint(0, output_size, batch_size).astype('int32')
tg_output = fun_output_tg(x_mat, y_mat) tg_output = fun_output_tg(x_mat, y_mat)
all_outputs = fun_output(x_mat) all_outputs = fun_output(x_mat)
...@@ -1730,23 +1730,23 @@ def test_h_softmax(): ...@@ -1730,23 +1730,23 @@ def test_h_softmax():
# Verifies that the outputs computed by fun_output_tg are the same as those # Verifies that the outputs computed by fun_output_tg are the same as those
# computed by fun_output. # computed by fun_output.
utt.assert_allclose( utt.assert_allclose(
all_outputs[numpy.arange(0, batch_size), y_mat], tg_output) all_outputs[np.arange(0, batch_size), y_mat], tg_output)
def test_elu(): def test_elu():
x = matrix('x') x = matrix('x')
seed = theano.tests.unittest_tools.fetch_seed() seed = theano.tests.unittest_tools.fetch_seed()
rng = numpy.random.RandomState(seed) rng = np.random.RandomState(seed)
X = rng.randn(20, 30).astype(config.floatX) X = rng.randn(20, 30).astype(config.floatX)
# test the base case, without custom alpha value # test the base case, without custom alpha value
y = elu(x).eval({x: X}) y = elu(x).eval({x: X})
utt.assert_allclose(y, numpy.where(X > 0, X, numpy.exp(X) - 1)) utt.assert_allclose(y, np.where(X > 0, X, np.exp(X) - 1))
# test for different constant alpha values # test for different constant alpha values
for alpha in 1.5, 2, -1, -1.5, -2: for alpha in 1.5, 2, -1, -1.5, -2:
y = elu(x, alpha).eval({x: X}) y = elu(x, alpha).eval({x: X})
utt.assert_allclose(y, numpy.where(X > 0, X, alpha * (numpy.exp(X) - 1))) utt.assert_allclose(y, np.where(X > 0, X, alpha * (np.exp(X) - 1)))
def test_binary_crossentropy_reshape(): def test_binary_crossentropy_reshape():
...@@ -1759,13 +1759,13 @@ def test_binary_crossentropy_reshape(): ...@@ -1759,13 +1759,13 @@ def test_binary_crossentropy_reshape():
# This only works when "specialize" options are included # This only works when "specialize" options are included
mode = theano.compile.get_default_mode().including('fast_run') mode = theano.compile.get_default_mode().including('fast_run')
fga = theano.function([a], ga, mode=mode) fga = theano.function([a], ga, mode=mode)
utt.assert_allclose(fga(numpy.array([[[[30.]]]], dtype=config.floatX)), utt.assert_allclose(fga(np.array([[[[30.]]]], dtype=config.floatX)),
numpy.zeros((1, 1, 1, 1), dtype=config.floatX)) np.zeros((1, 1, 1, 1), dtype=config.floatX))
SoftsignTester = makeBroadcastTester( SoftsignTester = makeBroadcastTester(
op=softsign, op=softsign,
expected=upcast_int8_nfunc(lambda inputs: check_floatX( expected=upcast_int8_nfunc(lambda inputs: check_floatX(
inputs, inputs / (1.0 + numpy.fabs(inputs)))), inputs, inputs / (1.0 + np.fabs(inputs)))),
good=_good_broadcast_unary_normal_float_no_complex, good=_good_broadcast_unary_normal_float_no_complex,
name='SoftsignTester', name='SoftsignTester',
) )
...@@ -1774,13 +1774,13 @@ SoftsignTester = makeBroadcastTester( ...@@ -1774,13 +1774,13 @@ SoftsignTester = makeBroadcastTester(
def test_confusion_matrix(): def test_confusion_matrix():
# Defining numpy implementation of confusion matrix # Defining numpy implementation of confusion matrix
def numpy_conf_mat(actual, pred): def numpy_conf_mat(actual, pred):
order = numpy.union1d(actual, pred) order = np.union1d(actual, pred)
colA = numpy.matrix(actual).T colA = np.matrix(actual).T
colP = numpy.matrix(pred).T colP = np.matrix(pred).T
oneHotA = colA.__eq__(order).astype('int64') oneHotA = colA.__eq__(order).astype('int64')
oneHotP = colP.__eq__(order).astype('int64') oneHotP = colP.__eq__(order).astype('int64')
conf_mat = numpy.dot(oneHotA.T, oneHotP) conf_mat = np.dot(oneHotA.T, oneHotP)
conf_mat = numpy.asarray(conf_mat) conf_mat = np.asarray(conf_mat)
return [conf_mat, order] return [conf_mat, order]
x = tensor.vector() x = tensor.vector()
...@@ -1790,8 +1790,8 @@ def test_confusion_matrix(): ...@@ -1790,8 +1790,8 @@ def test_confusion_matrix():
[[2, 0, 2, 2, 0, 1], [0, 0, 2, 2, 0, 2]]] [[2, 0, 2, 2, 0, 1], [0, 0, 2, 2, 0, 2]]]
for case in list_inputs: for case in list_inputs:
a = numpy.asarray(case[0]) a = np.asarray(case[0])
b = numpy.asarray(case[1]) b = np.asarray(case[1])
out_exp = numpy_conf_mat(a, b) out_exp = numpy_conf_mat(a, b)
outs = f(case[0], case[1]) outs = f(case[0], case[1])
for exp, out in zip(out_exp, outs): for exp, out in zip(out_exp, outs):
......
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