Skip to content

P
pytensor
提交 3d40556c authored 作者: Frederic's avatar Frederic
浏览文件
操作

pep8

上级 d2d8f01b
隐藏空白字符变更
内嵌 并排
正在显示 包含 62 行增加 和 61 行删除
theano/tests/test_tutorial.py
浏览文件 @ 3d40556c
...@@ -1145,26 +1145,26 @@ class T_scan(unittest.TestCase): ...@@ -1145,26 +1145,26 @@ class T_scan(unittest.TestCase):
## Any change you do here also add it to the tutorial ! ## Any change you do here also add it to the tutorial !
def test_elemwise(self): def test_elemwise(self):
# defining the tensor variables # defining the tensor variables
X = T.matrix("X") X = T.matrix("X")
W = T.matrix("W") W = T.matrix("W")
b_sym = T.vector("b_sym") b_sym = T.vector("b_sym")
results, updates = theano.scan(lambda v:T.tanh(T.dot(v,W)+b_sym), results, updates = theano.scan(lambda v: T.tanh(T.dot(v, W) + b_sym),
sequences=X) sequences=X)
compute_elementwise = theano.function(inputs = [X, W, b_sym], compute_elementwise = theano.function(inputs=[X, W, b_sym],
outputs=[results]) outputs=[results])
# test values # test values
x = numpy.eye(2) x = numpy.eye(2)
w = numpy.ones((2,2)) w = numpy.ones((2, 2))
b = numpy.ones((2)) b = numpy.ones((2))
b[1] = 2 b[1] = 2
print "Scan results:", compute_elementwise(x, w, b)[0] print "Scan results:", compute_elementwise(x, w, b)[0]
# comparison with numpy # comparison with numpy
print "Numpy results:", numpy.tanh(x.dot(w) + b) print "Numpy results:", numpy.tanh(x.dot(w) + b)
def test_sequence(self): def test_sequence(self):
# define tensor variables # define tensor variables
...@@ -1176,30 +1176,31 @@ class T_scan(unittest.TestCase): ...@@ -1176,30 +1176,31 @@ class T_scan(unittest.TestCase):
V = T.matrix("V") V = T.matrix("V")
P = T.matrix("P") P = T.matrix("P")
results, updates = theano.scan(lambda y,p,x_tm1:T.tanh(T.dot(x_tm1,W) + results, updates = theano.scan(
T.dot(y,U)+T.dot(p,V)), lambda y, p, x_tm1: T.tanh(T.dot(x_tm1, W) +
sequences=[Y,P[::-1]], outputs_info=[X]) T.dot(y, U) + T.dot(p, V)),
sequences=[Y, P[::-1]], outputs_info=[X])
compute_seq = theano.function(inputs = [X, W, Y, U, P, V], compute_seq = theano.function(inputs=[X, W, Y, U, P, V],
outputs=[results]) outputs=[results])
# test values # test values
x = numpy.zeros((2)) x = numpy.zeros((2))
x[1] = 1 x[1] = 1
w = numpy.ones((2,2)) w = numpy.ones((2, 2))
y = numpy.ones((5,2)) y = numpy.ones((5, 2))
y[0,:] = -3 y[0, :] = -3
u = numpy.ones((2,2)) u = numpy.ones((2, 2))
p = numpy.ones((5,2)) p = numpy.ones((5, 2))
p[0,:] = 3 p[0, :] = 3
v = numpy.ones((2,2)) v = numpy.ones((2, 2))
print "Scan results", compute_seq(x,w,y,u,p,v)[0] print "Scan results", compute_seq(x, w, y, u, p, v)[0]
# comparison with numpy # comparison with numpy
x_res = numpy.zeros((5,2)) x_res = numpy.zeros((5, 2))
x_res[0] = numpy.tanh(x.dot(w) + y[0].dot(u) + p[4].dot(v)) x_res[0] = numpy.tanh(x.dot(w) + y[0].dot(u) + p[4].dot(v))
for i in range(1,5): for i in range(1, 5):
x_res[i] = numpy.tanh(x_res[i-1].dot(w) + x_res[i] = numpy.tanh(x_res[i-1].dot(w) +
y[i].dot(u) + p[4-i].dot(v)) y[i].dot(u) + p[4-i].dot(v))
...@@ -1208,16 +1209,16 @@ class T_scan(unittest.TestCase): ...@@ -1208,16 +1209,16 @@ class T_scan(unittest.TestCase):
def test_norm(self): def test_norm(self):
# define tensor variable # define tensor variable
X = T.matrix("X") X = T.matrix("X")
results, updates = theano.scan(lambda x_i:T.sqrt((x_i**2).sum()), results, updates = theano.scan(lambda x_i: T.sqrt((x_i**2).sum()),
sequences=[X]) sequences=[X])
compute_norm_lines = theano.function(inputs = [X], outputs=[results]) compute_norm_lines = theano.function(inputs=[X], outputs=[results])
results, updates = theano.scan(lambda x_i:T.sqrt((x_i**2).sum()), results, updates = theano.scan(lambda x_i: T.sqrt((x_i**2).sum()),
sequences=[X.T]) sequences=[X.T])
compute_norm_cols = theano.function(inputs = [X], outputs=[results]) compute_norm_cols = theano.function(inputs=[X], outputs=[results])
# test value # test value
x = numpy.diag(numpy.arange(1,6),1) x = numpy.diag(numpy.arange(1, 6), 1)
print "Scan results:", compute_norm_lines(x)[0], \ print "Scan results:", compute_norm_lines(x)[0], \
compute_norm_cols(x)[0] compute_norm_cols(x)[0]
...@@ -1228,15 +1229,15 @@ class T_scan(unittest.TestCase): ...@@ -1228,15 +1229,15 @@ class T_scan(unittest.TestCase):
def test_trace(self): def test_trace(self):
# define tensor variable # define tensor variable
X = T.matrix("X") X = T.matrix("X")
results, updates = theano.scan(lambda i, j, t_f:T.cast(X[i,j] + results, updates = theano.scan(lambda i, j, t_f: T.cast(X[i,j] +
t_f, theano.config.floatX), t_f, theano.config.floatX),
sequences=[T.arange(X.shape[0]), sequences=[T.arange(X.shape[0]),
T.arange(X.shape[1])], T.arange(X.shape[1])],
outputs_info=numpy.asarray( outputs_info=numpy.asarray(
0., dtype=theano.config.floatX)) 0., dtype=theano.config.floatX))
result = results[-1] result = results[-1]
compute_trace = theano.function(inputs = [X], outputs=[result]) compute_trace = theano.function(inputs=[X], outputs=[result])
# test value # test value
x = numpy.eye(5) x = numpy.eye(5)
...@@ -1256,33 +1257,33 @@ class T_scan(unittest.TestCase): ...@@ -1256,33 +1257,33 @@ class T_scan(unittest.TestCase):
n_sym = T.iscalar("n_sym") n_sym = T.iscalar("n_sym")
results, updates = theano.scan( results, updates = theano.scan(
lambda x_tm2,x_tm1:T.dot(x_tm2,U) + T.dot(x_tm1,V) + T.tanh(T.dot(x_tm1,W) + b_sym), lambda x_tm2,x_tm1: T.dot(x_tm2,U) + T.dot(x_tm1,V) + T.tanh(T.dot(x_tm1,W) + b_sym),
n_steps=n_sym, n_steps=n_sym,
outputs_info=[dict(initial = X, taps = [-2,-1])]) outputs_info=[dict(initial=X, taps=[-2, -1])])
compute_seq2 = theano.function(inputs = [X, U, V, W, b_sym, n_sym], compute_seq2 = theano.function(inputs=[X, U, V, W, b_sym, n_sym],
outputs=[results]) outputs=[results])
# test values # test values
x = numpy.zeros((2,2)) x = numpy.zeros((2, 2))
# the initial value must be able to return x[-2] # the initial value must be able to return x[-2]
x[1,1] = 1 x[1, 1] = 1
w = 0.5*numpy.ones((2,2)) w = 0.5 * numpy.ones((2, 2))
u = 0.5*(numpy.ones((2,2))-numpy.eye(2)) u = 0.5 * (numpy.ones((2, 2)) - numpy.eye(2))
v = 0.5*numpy.ones((2,2)) v = 0.5 * numpy.ones((2, 2))
n = 10 n = 10
b = numpy.ones((2)) b = numpy.ones((2))
print "Scan results:", compute_seq2(x,u,v,w,b,n) print "Scan results:", compute_seq2(x, u, v, w, b, n)
# comparison with numpy # comparison with numpy
x_res = numpy.zeros((10,2)) x_res = numpy.zeros((10, 2))
x_res[0] = x[0].dot(u) + x[1].dot(v) + numpy.tanh(x[1].dot(w) + b) x_res[0] = x[0].dot(u) + x[1].dot(v) + numpy.tanh(x[1].dot(w) + b)
x_res[1] = x[1].dot(u) + x_res[0].dot(v) \ x_res[1] = x[1].dot(u) + x_res[0].dot(v) \
+ numpy.tanh(x_res[0].dot(w) + b) + numpy.tanh(x_res[0].dot(w) + b)
x_res[2] = x_res[0].dot(u) + x_res[1].dot(v) \ x_res[2] = x_res[0].dot(u) + x_res[1].dot(v) \
+ numpy.tanh(x_res[1].dot(w) + b) + numpy.tanh(x_res[1].dot(w) + b)
for i in range(2,10): for i in range(2, 10):
x_res[i] = (x_res[i-2].dot(u) + x_res[i-1].dot(v) + x_res[i] = (x_res[i-2].dot(u) + x_res[i-1].dot(v) +
numpy.tanh(x_res[i-1].dot(w) + b)) numpy.tanh(x_res[i-1].dot(w) + b))
...@@ -1292,17 +1293,17 @@ class T_scan(unittest.TestCase): ...@@ -1292,17 +1293,17 @@ class T_scan(unittest.TestCase):
# define tensor variables # define tensor variables
v = T.vector() v = T.vector()
A = T.matrix() A = T.matrix()
y = T.tanh(T.dot(v,A)) y = T.tanh(T.dot(v, A))
results, updates = theano.scan(lambda i:T.grad(y[i], v), results, updates = theano.scan(lambda i: T.grad(y[i], v),
sequences = [T.arange(y.shape[0])]) sequences=[T.arange(y.shape[0])])
compute_jac_t = theano.function([A,v], [results], compute_jac_t = theano.function([A, v], [results],
allow_input_downcast = True) # shape (d_out, d_in) allow_input_downcast=True) # shape (d_out, d_in)
# test values # test values
x = numpy.eye(5)[0] x = numpy.eye(5)[0]
w = numpy.eye(5,3) w = numpy.eye(5, 3)
w[2] = numpy.ones((3)) w[2] = numpy.ones((3))
print "Scan results:", compute_jac_t(w,x)[0] print "Scan results:", compute_jac_t(w, x)[0]
# compare with numpy # compare with numpy
print "Numpy results:", ((1 - numpy.tanh(x.dot(w))**2)*w).T print "Numpy results:", ((1 - numpy.tanh(x.dot(w))**2)*w).T
...@@ -1312,9 +1313,9 @@ class T_scan(unittest.TestCase): ...@@ -1312,9 +1313,9 @@ class T_scan(unittest.TestCase):
k = theano.shared(0) k = theano.shared(0)
n_sym = T.iscalar("n_sym") n_sym = T.iscalar("n_sym")
results, updates = theano.scan(lambda:{k:(k+1)}, n_steps=n_sym) results, updates = theano.scan(lambda: {k: (k + 1)}, n_steps=n_sym)
accumulator = theano.function([n_sym], [], updates=updates, accumulator = theano.function([n_sym], [], updates=updates,
allow_input_downcast = True) allow_input_downcast=True)
print "Before 5 steps:", k.get_value() print "Before 5 steps:", k.get_value()
accumulator(5) accumulator(5)
...@@ -1328,11 +1329,11 @@ class T_scan(unittest.TestCase): ...@@ -1328,11 +1329,11 @@ class T_scan(unittest.TestCase):
# define shared random stream # define shared random stream
trng = T.shared_randomstreams.RandomStreams(1234) trng = T.shared_randomstreams.RandomStreams(1234)
d=trng.binomial(size=W[1].shape) d = trng.binomial(size=W[1].shape)
results, updates = theano.scan(lambda v:T.tanh(T.dot(v,W) + b_sym)*d, results, updates = theano.scan(lambda v: T.tanh(T.dot(v, W) + b_sym) * d,
sequences=X) sequences=X)
compute_with_bnoise = theano.function(inputs = [X, W, b_sym], compute_with_bnoise = theano.function(inputs=[X, W, b_sym],
outputs=[results], outputs=[results],
updates=updates, updates=updates,
allow_input_downcast = True) allow_input_downcast = 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