Skip to content

P
pytensor
提交 7e05e00e authored 作者: amrithasuresh's avatar amrithasuresh
浏览文件
操作

1. Updated numpy as np

2. Fixed indentation
上级 3c237906
显示空白字符变更
内嵌 并排
正在显示 包含 145 行增加 和 145 行删除
theano/tensor/tests/test_raw_random.py
浏览文件 @ 7e05e00e
from __future__ import absolute_import, print_function, division from __future__ import absolute_import, print_function, division
import numpy import numpy as np
import pickle import pickle
from theano.tests import unittest_tools as utt from theano.tests import unittest_tools as utt
...@@ -19,7 +19,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -19,7 +19,7 @@ class T_random_function(utt.InferShapeTester):
utt.seed_rng() utt.seed_rng()
def test_basic_usage(self): def test_basic_usage(self):
rf = RandomFunction(numpy.random.RandomState.uniform, tensor.dvector) rf = RandomFunction(np.random.RandomState.uniform, tensor.dvector)
assert not rf.inplace assert not rf.inplace
assert getattr(rf, 'destroy_map', {}) == {} assert getattr(rf, 'destroy_map', {}) == {}
...@@ -33,23 +33,23 @@ class T_random_function(utt.InferShapeTester): ...@@ -33,23 +33,23 @@ class T_random_function(utt.InferShapeTester):
f = compile.function([rng_R], out) f = compile.function([rng_R], out)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
f_0 = f(rng_state0) f_0 = f(rng_state0)
f_1 = f(rng_state0) f_1 = f(rng_state0)
assert numpy.all(f_0 == f_1) assert np.all(f_0 == f_1)
def test_inplace_norun(self): def test_inplace_norun(self):
rf = RandomFunction(numpy.random.RandomState.uniform, tensor.dvector, rf = RandomFunction(np.random.RandomState.uniform, tensor.dvector,
inplace=True) inplace=True)
assert rf.inplace assert rf.inplace
assert getattr(rf, 'destroy_map', {}) != {} assert getattr(rf, 'destroy_map', {}) != {}
def test_args(self): def test_args(self):
"""Test that arguments to RandomFunction are honored""" """Test that arguments to RandomFunction are honored"""
rf2 = RandomFunction(numpy.random.RandomState.uniform, tensor.dvector) rf2 = RandomFunction(np.random.RandomState.uniform, tensor.dvector)
rf4 = RandomFunction(numpy.random.RandomState.uniform, tensor.dvector, rf4 = RandomFunction(np.random.RandomState.uniform, tensor.dvector,
inplace=True) inplace=True)
rng_R = random_state_type() rng_R = random_state_type()
...@@ -64,7 +64,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -64,7 +64,7 @@ class T_random_function(utt.InferShapeTester):
# be maintained by post_r4 # be maintained by post_r4
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r4, update=post_r4,
mutable=True)], mutable=True)],
[out2, out4, out2_4, out2_4_4], [out2, out4, out2_4, out2_4_4],
...@@ -84,19 +84,19 @@ class T_random_function(utt.InferShapeTester): ...@@ -84,19 +84,19 @@ class T_random_function(utt.InferShapeTester):
# print f2_4_4b # print f2_4_4b
# setting bounds is same as multiplying by 2 # setting bounds is same as multiplying by 2
assert numpy.allclose(f2 * 2, f4), (f2, f4) assert np.allclose(f2 * 2, f4), (f2, f4)
# retrieving from non-inplace generator # retrieving from non-inplace generator
# is same as inplace one for first call # is same as inplace one for first call
assert numpy.allclose(f2_4_4, f4), (f2_4_4, f4) assert np.allclose(f2_4_4, f4), (f2_4_4, f4)
# f4 changes from call to call, that the update has worked # f4 changes from call to call, that the update has worked
assert not numpy.allclose(f4, f4b), (f4, f4b) assert not np.allclose(f4, f4b), (f4, f4b)
def test_inplace_optimization(self): def test_inplace_optimization(self):
"""Test that FAST_RUN includes the random_make_inplace optimization""" """Test that FAST_RUN includes the random_make_inplace optimization"""
#inplace = False #inplace = False
rf2 = RandomFunction(numpy.random.RandomState.uniform, tensor.dvector) rf2 = RandomFunction(np.random.RandomState.uniform, tensor.dvector)
rng_R = random_state_type() rng_R = random_state_type()
# If calling RandomFunction directly, all args have to be specified, # If calling RandomFunction directly, all args have to be specified,
...@@ -105,7 +105,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -105,7 +105,7 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r2, update=post_r2,
mutable=True)], mutable=True)],
out2, out2,
...@@ -121,7 +121,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -121,7 +121,7 @@ class T_random_function(utt.InferShapeTester):
val1 = f() val1 = f()
assert id0 == id(f[rng_R]) assert id0 == id(f[rng_R])
assert not numpy.allclose(val0, val1) assert not np.allclose(val0, val1)
def test_no_inplace(self): def test_no_inplace(self):
"""Test that when not running inplace, the RandomState is """Test that when not running inplace, the RandomState is
...@@ -131,10 +131,10 @@ class T_random_function(utt.InferShapeTester): ...@@ -131,10 +131,10 @@ class T_random_function(utt.InferShapeTester):
post_r, out = rf(rng_R, (3,), 0., 1.) post_r, out = rf(rng_R, (3,), 0., 1.)
f = compile.function([rng_R], [post_r, out]) f = compile.function([rng_R], [post_r, out])
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
rng0, val0 = f(rng) rng0, val0 = f(rng)
rng_ = numpy.random.RandomState(utt.fetch_seed()) rng_ = np.random.RandomState(utt.fetch_seed())
# rng should still be in a fresh state # rng should still be in a fresh state
self.assertTrue(rng_R.type.values_eq(rng, rng_)) self.assertTrue(rng_R.type.values_eq(rng, rng_))
# rng0 should be in an updated state # rng0 should be in an updated state
...@@ -178,7 +178,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -178,7 +178,7 @@ class T_random_function(utt.InferShapeTester):
f_ok = compile.function( f_ok = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_out2_4_4, update=post_out2_4_4,
mutable=True)], mutable=True)],
[out4, out1_4, out2_4_4], [out4, out1_4, out2_4_4],
...@@ -188,8 +188,8 @@ class T_random_function(utt.InferShapeTester): ...@@ -188,8 +188,8 @@ class T_random_function(utt.InferShapeTester):
o4, o1_4, o2_4_4 = f_ok() o4, o1_4, o2_4_4 = f_ok()
# Check the sanity of the answers # Check the sanity of the answers
self.assertTrue(numpy.allclose(o4, o1_4)) self.assertTrue(np.allclose(o4, o1_4))
self.assertTrue(numpy.allclose(o4, o2_4_4[0])) self.assertTrue(np.allclose(o4, o2_4_4[0]))
def test_random_function_noshape_args(self): def test_random_function_noshape_args(self):
'''Test if random_function helper works with args but without shape''' '''Test if random_function helper works with args but without shape'''
...@@ -199,7 +199,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -199,7 +199,7 @@ class T_random_function(utt.InferShapeTester):
post_out, out = uniform(rng_R, size=None, ndim=2) post_out, out = uniform(rng_R, size=None, ndim=2)
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_out, update=post_out,
mutable=True)], mutable=True)],
[out], [out],
...@@ -219,7 +219,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -219,7 +219,7 @@ class T_random_function(utt.InferShapeTester):
[low, [low,
high, high,
compile.In(rng_R, compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_out2, update=post_out2,
mutable=True)], mutable=True)],
[out2], [out2],
...@@ -242,7 +242,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -242,7 +242,7 @@ class T_random_function(utt.InferShapeTester):
# If using numpy's uniform distribution, ndim_added should be 0, # If using numpy's uniform distribution, ndim_added should be 0,
# because the shape provided as argument is the output shape. # because the shape provided as argument is the output shape.
# Specifying a different ndim_added will change the Op's output ndim, # Specifying a different ndim_added will change the Op's output ndim,
# so numpy.uniform will produce a result of incorrect shape, # so np.uniform will produce a result of incorrect shape,
# and a ValueError should be raised. # and a ValueError should be raised.
def ndim_added_deco(ndim_added): def ndim_added_deco(ndim_added):
def randomfunction(random_state, size=(), low=0.0, high=0.0, def randomfunction(random_state, size=(), low=0.0, high=0.0,
...@@ -282,27 +282,27 @@ class T_random_function(utt.InferShapeTester): ...@@ -282,27 +282,27 @@ class T_random_function(utt.InferShapeTester):
f11 = compile.function( f11 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=p_uni11, mutable=True)], update=p_uni11, mutable=True)],
[uni11], accept_inplace=True) [uni11], accept_inplace=True)
f12 = compile.function( f12 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=p_uni12, mutable=True)], update=p_uni12, mutable=True)],
[uni12], accept_inplace=True) [uni12], accept_inplace=True)
fm11 = compile.function( fm11 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=p_unim11, mutable=True)], update=p_unim11, mutable=True)],
[unim11], accept_inplace=True) [unim11], accept_inplace=True)
fm12 = compile.function( fm12 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=p_unim12, mutable=True)], update=p_unim12, mutable=True)],
[unim12], accept_inplace=True) [unim12], accept_inplace=True)
f0 = compile.function( f0 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=p_uni02, mutable=True)], update=p_uni02, mutable=True)],
[uni01, uni02], accept_inplace=True) [uni01, uni02], accept_inplace=True)
self.assertRaises(ValueError, f11) self.assertRaises(ValueError, f11)
...@@ -310,7 +310,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -310,7 +310,7 @@ class T_random_function(utt.InferShapeTester):
self.assertRaises(ValueError, fm11) self.assertRaises(ValueError, fm11)
self.assertRaises(ValueError, fm12) self.assertRaises(ValueError, fm12)
u01, u02 = f0() u01, u02 = f0()
self.assertTrue(numpy.allclose(u01, u02[0])) self.assertTrue(np.allclose(u01, u02[0]))
def test_uniform(self): def test_uniform(self):
"""Test that raw_random.uniform generates the same results as numpy.""" """Test that raw_random.uniform generates the same results as numpy."""
...@@ -321,17 +321,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -321,17 +321,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.uniform(-2.0, 2.0, size=(4,)) numpy_val0 = numpy_rng.uniform(-2.0, 2.0, size=(4,))
numpy_val1 = numpy_rng.uniform(-2.0, 2.0, size=(4,)) numpy_val1 = numpy_rng.uniform(-2.0, 2.0, size=(4,))
self.assertTrue(numpy.allclose(val0, numpy_val0)) self.assertTrue(np.allclose(val0, numpy_val0))
self.assertTrue(numpy.allclose(val1, numpy_val1)) self.assertTrue(np.allclose(val1, numpy_val1))
def test_binomial(self): def test_binomial(self):
"""Test that raw_random.binomial generates the same results """Test that raw_random.binomial generates the same results
...@@ -344,17 +344,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -344,17 +344,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[bin], accept_inplace=True) [bin], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.binomial(5, 0.8, size=(7, 12)) numpy_val0 = numpy_rng.binomial(5, 0.8, size=(7, 12))
numpy_val1 = numpy_rng.binomial(5, 0.8, size=(7, 12)) numpy_val1 = numpy_rng.binomial(5, 0.8, size=(7, 12))
self.assertTrue(numpy.all(val0 == numpy_val0)) self.assertTrue(np.all(val0 == numpy_val0))
self.assertTrue(numpy.all(val1 == numpy_val1)) self.assertTrue(np.all(val1 == numpy_val1))
def test_normal(self): def test_normal(self):
"""Test that raw_random.normal generates the same results as numpy.""" """Test that raw_random.normal generates the same results as numpy."""
...@@ -365,17 +365,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -365,17 +365,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.normal(4.0, 2.0, size=(2, 3)) numpy_val0 = numpy_rng.normal(4.0, 2.0, size=(2, 3))
numpy_val1 = numpy_rng.normal(4.0, 2.0, size=(2, 3)) numpy_val1 = numpy_rng.normal(4.0, 2.0, size=(2, 3))
self.assertTrue(numpy.allclose(val0, numpy_val0)) self.assertTrue(np.allclose(val0, numpy_val0))
self.assertTrue(numpy.allclose(val1, numpy_val1)) self.assertTrue(np.allclose(val1, numpy_val1))
def test_random_integers(self): def test_random_integers(self):
# Test that raw_random.random_integers generates the same # Test that raw_random.random_integers generates the same
...@@ -390,17 +390,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -390,17 +390,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.randint(-3, 17, size=(11, 8)) numpy_val0 = numpy_rng.randint(-3, 17, size=(11, 8))
numpy_val1 = numpy_rng.randint(-3, 17, size=(11, 8)) numpy_val1 = numpy_rng.randint(-3, 17, size=(11, 8))
self.assertTrue(numpy.allclose(val0, numpy_val0)) self.assertTrue(np.allclose(val0, numpy_val0))
self.assertTrue(numpy.allclose(val1, numpy_val1)) self.assertTrue(np.allclose(val1, numpy_val1))
def test_permutation_helper(self): def test_permutation_helper(self):
"""Test that raw_random.permutation_helper generates the same """Test that raw_random.permutation_helper generates the same
...@@ -418,21 +418,21 @@ class T_random_function(utt.InferShapeTester): ...@@ -418,21 +418,21 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
# numpy_rng.permutation outputs one vector at a time, # numpy_rng.permutation outputs one vector at a time,
# so we call it iteratively to generate all the samples. # so we call it iteratively to generate all the samples.
numpy_val0 = numpy.asarray([numpy_rng.permutation(8) numpy_val0 = np.asarray([numpy_rng.permutation(8)
for i in range(7)]) for i in range(7)])
numpy_val1 = numpy.asarray([numpy_rng.permutation(8) numpy_val1 = np.asarray([numpy_rng.permutation(8)
for i in range(7)]) for i in range(7)])
self.assertTrue(numpy.all(val0 == numpy_val0)) self.assertTrue(np.all(val0 == numpy_val0))
self.assertTrue(numpy.all(val1 == numpy_val1)) self.assertTrue(np.all(val1 == numpy_val1))
# This call lacks "ndim_added=1", so ndim_added defaults to 0. # This call lacks "ndim_added=1", so ndim_added defaults to 0.
# A ValueError should be raised. # A ValueError should be raised.
...@@ -440,7 +440,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -440,7 +440,7 @@ class T_random_function(utt.InferShapeTester):
post_r0, out0 = rf0(rng_R, (7,), 8) post_r0, out0 = rf0(rng_R, (7,), 8)
f0 = compile.function( f0 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r0, mutable=True)], update=post_r0, mutable=True)],
[out0], accept_inplace=True) [out0], accept_inplace=True)
self.assertRaises(ValueError, f0) self.assertRaises(ValueError, f0)
...@@ -451,7 +451,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -451,7 +451,7 @@ class T_random_function(utt.InferShapeTester):
post_r2, out2 = rf2(rng_R, (7,), 8) post_r2, out2 = rf2(rng_R, (7,), 8)
f2 = compile.function( f2 = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r2, mutable=True)], update=post_r2, mutable=True)],
[out2], accept_inplace=True) [out2], accept_inplace=True)
self.assertRaises(ValueError, f2) self.assertRaises(ValueError, f2)
...@@ -467,17 +467,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -467,17 +467,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.choice(10, (11, 8), True, None) numpy_val0 = numpy_rng.choice(10, (11, 8), True, None)
numpy_val1 = numpy_rng.choice(10, (11, 8), True, None) numpy_val1 = numpy_rng.choice(10, (11, 8), True, None)
self.assertTrue(numpy.allclose(val0, numpy_val0)) self.assertTrue(np.allclose(val0, numpy_val0))
self.assertTrue(numpy.allclose(val1, numpy_val1)) self.assertTrue(np.allclose(val1, numpy_val1))
def test_poisson(self): def test_poisson(self):
"""Test that raw_random.poisson generates the same """Test that raw_random.poisson generates the same
...@@ -490,17 +490,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -490,17 +490,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy_rng.poisson(5, size=(11, 8)) numpy_val0 = numpy_rng.poisson(5, size=(11, 8))
numpy_val1 = numpy_rng.poisson(5, size=(11, 8)) numpy_val1 = numpy_rng.poisson(5, size=(11, 8))
self.assertTrue(numpy.allclose(val0, numpy_val0)) self.assertTrue(np.allclose(val0, numpy_val0))
self.assertTrue(numpy.allclose(val1, numpy_val1)) self.assertTrue(np.allclose(val1, numpy_val1))
def test_permutation(self): def test_permutation(self):
"""Test that raw_random.permutation generates the same """Test that raw_random.permutation generates the same
...@@ -509,33 +509,33 @@ class T_random_function(utt.InferShapeTester): ...@@ -509,33 +509,33 @@ class T_random_function(utt.InferShapeTester):
post_r, out = permutation(rng_R, size=(9,), n=6) post_r, out = permutation(rng_R, size=(9,), n=6)
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Check over two calls to see if the random state is correctly updated. # Check over two calls to see if the random state is correctly updated.
# numpy_rng.permutation outputs one vector at a time, # numpy_rng.permutation outputs one vector at a time,
# so we call it iteratively to generate all the samples. # so we call it iteratively to generate all the samples.
val0 = f() val0 = f()
val1 = f() val1 = f()
numpy_val0 = numpy.asarray([numpy_rng.permutation(6) numpy_val0 = np.asarray([numpy_rng.permutation(6)
for i in range(9)]) for i in range(9)])
numpy_val1 = numpy.asarray([numpy_rng.permutation(6) numpy_val1 = np.asarray([numpy_rng.permutation(6)
for i in range(9)]) for i in range(9)])
self.assertTrue(numpy.all(val0 == numpy_val0)) self.assertTrue(np.all(val0 == numpy_val0))
self.assertTrue(numpy.all(val1 == numpy_val1)) self.assertTrue(np.all(val1 == numpy_val1))
# Test that we can generate a list: have size=None or (). # Test that we can generate a list: have size=None or ().
for ndim in [1, None]: for ndim in [1, None]:
post_r, out = permutation(rng_R, n=10, size=None, ndim=ndim) post_r, out = permutation(rng_R, n=10, size=None, ndim=ndim)
inp = compile.In(rng_R, inp = compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True) update=post_r, mutable=True)
f = theano.function([inp], out) f = theano.function([inp], out)
o = f() o = f()
assert o.shape == (10,) assert o.shape == (10,)
assert (numpy.sort(o) == numpy.arange(10)).all() assert (np.sort(o) == np.arange(10)).all()
# Wrong number of dimensions asked # Wrong number of dimensions asked
self.assertRaises(TypeError, permutation, rng_R, size=None, ndim=2) self.assertRaises(TypeError, permutation, rng_R, size=None, ndim=2)
...@@ -548,17 +548,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -548,17 +548,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function( f = compile.function(
[compile.In(rng_R, [compile.In(rng_R,
value=numpy.random.RandomState(utt.fetch_seed()), value=np.random.RandomState(utt.fetch_seed()),
update=post_r, mutable=True)], update=post_r, mutable=True)],
[out], accept_inplace=True) [out], accept_inplace=True)
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
val0, = f() val0, = f()
val1, = f() val1, = f()
numpy_val0 = numpy_rng.multinomial(6, [0.2] * 5, (7, 3)) numpy_val0 = numpy_rng.multinomial(6, [0.2] * 5, (7, 3))
numpy_val1 = numpy_rng.multinomial(6, [0.2] * 5, (7, 3)) numpy_val1 = numpy_rng.multinomial(6, [0.2] * 5, (7, 3))
self.assertTrue(numpy.all(val0 == numpy_val0)) self.assertTrue(np.all(val0 == numpy_val0))
self.assertTrue(numpy.all(val1 == numpy_val1)) self.assertTrue(np.all(val1 == numpy_val1))
self.assertTrue(val0.shape == (7, 3, 5)) self.assertTrue(val0.shape == (7, 3, 5))
self.assertTrue(val1.shape == (7, 3, 5)) self.assertTrue(val1.shape == (7, 3, 5))
...@@ -568,7 +568,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -568,7 +568,7 @@ class T_random_function(utt.InferShapeTester):
shape = tensor.lvector() shape = tensor.lvector()
post_r, out = uniform(rng_R, shape, ndim=2) post_r, out = uniform(rng_R, shape, ndim=2)
f = compile.function([rng_R, shape], out) f = compile.function([rng_R, shape], out)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
assert f(rng_state0, [2, 3]).shape == (2, 3) assert f(rng_state0, [2, 3]).shape == (2, 3)
assert f(rng_state0, [4, 8]).shape == (4, 8) assert f(rng_state0, [4, 8]).shape == (4, 8)
...@@ -583,7 +583,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -583,7 +583,7 @@ class T_random_function(utt.InferShapeTester):
shape = (shape0, 3) shape = (shape0, 3)
post_r, u = uniform(rng_R, size=shape, ndim=2) post_r, u = uniform(rng_R, size=shape, ndim=2)
f = compile.function([rng_R, shape0], u) f = compile.function([rng_R, shape0], u)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
assert f(rng_state0, 2).shape == (2, 3) assert f(rng_state0, 2).shape == (2, 3)
assert f(rng_state0, 8).shape == (8, 3) assert f(rng_state0, 8).shape == (8, 3)
...@@ -601,7 +601,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -601,7 +601,7 @@ class T_random_function(utt.InferShapeTester):
post_r, u = uniform(rng_R, size=shape, ndim=2) post_r, u = uniform(rng_R, size=shape, ndim=2)
assert u.broadcastable == (False, True) assert u.broadcastable == (False, True)
f = compile.function([rng_R, shape0], u) f = compile.function([rng_R, shape0], u)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
assert f(rng_state0, 2).shape == (2, 1) assert f(rng_state0, 2).shape == (2, 1)
assert f(rng_state0, 8).shape == (8, 1) assert f(rng_state0, 8).shape == (8, 1)
...@@ -617,25 +617,25 @@ class T_random_function(utt.InferShapeTester): ...@@ -617,25 +617,25 @@ class T_random_function(utt.InferShapeTester):
post_r, out = uniform(rng_R) post_r, out = uniform(rng_R)
f = compile.function([rng_R], [post_r, out], accept_inplace=True) f = compile.function([rng_R], [post_r, out], accept_inplace=True)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
post0, val0 = f(rng_state0) post0, val0 = f(rng_state0)
post1, val1 = f(post0) post1, val1 = f(post0)
numpy_val0 = numpy.asarray(numpy_rng.uniform(), numpy_val0 = np.asarray(numpy_rng.uniform(),
dtype=theano.config.floatX) dtype=theano.config.floatX)
numpy_val1 = numpy.asarray(numpy_rng.uniform(), numpy_val1 = np.asarray(numpy_rng.uniform(),
dtype=theano.config.floatX) dtype=theano.config.floatX)
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
post_r, out = multinomial(rng_R) post_r, out = multinomial(rng_R)
g = compile.function([rng_R], [post_r, out], accept_inplace=True) g = compile.function([rng_R], [post_r, out], accept_inplace=True)
post2, val2 = g(post1) post2, val2 = g(post1)
numpy_val2 = numpy.asarray(numpy_rng.multinomial(n=1, pvals=[.5, .5]), numpy_val2 = np.asarray(numpy_rng.multinomial(n=1, pvals=[.5, .5]),
dtype=theano.config.floatX) dtype=theano.config.floatX)
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
def test_vector_arguments(self): def test_vector_arguments(self):
rng_R = random_state_type() rng_R = random_state_type()
...@@ -645,17 +645,17 @@ class T_random_function(utt.InferShapeTester): ...@@ -645,17 +645,17 @@ class T_random_function(utt.InferShapeTester):
f = compile.function([rng_R, low], [post_r, out], accept_inplace=True) f = compile.function([rng_R, low], [post_r, out], accept_inplace=True)
def as_floatX(thing): def as_floatX(thing):
return numpy.asarray(thing, dtype=theano.config.floatX) return np.asarray(thing, dtype=theano.config.floatX)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
post0, val0 = f(rng_state0, [-5, .5, 0, 1]) post0, val0 = f(rng_state0, [-5, .5, 0, 1])
post1, val1 = f(post0, as_floatX([.9])) post1, val1 = f(post0, as_floatX([.9]))
numpy_val0 = as_floatX(numpy_rng.uniform(low=[-5, .5, 0, 1], high=1)) numpy_val0 = as_floatX(numpy_rng.uniform(low=[-5, .5, 0, 1], high=1))
numpy_val1 = as_floatX(numpy_rng.uniform(low=as_floatX([.9]), high=1)) numpy_val1 = as_floatX(numpy_rng.uniform(low=as_floatX([.9]), high=1))
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
high = tensor.vector() high = tensor.vector()
post_rb, outb = uniform(rng_R, low=low, high=high) post_rb, outb = uniform(rng_R, low=low, high=high)
...@@ -667,8 +667,8 @@ class T_random_function(utt.InferShapeTester): ...@@ -667,8 +667,8 @@ class T_random_function(utt.InferShapeTester):
post1b, val1b = fb(post0b, [-4.], [-1]) post1b, val1b = fb(post0b, [-4.], [-1])
numpy_val0b = as_floatX(numpy_rng.uniform(low=[-4., -2], high=[-1, 0])) numpy_val0b = as_floatX(numpy_rng.uniform(low=[-4., -2], high=[-1, 0]))
numpy_val1b = as_floatX(numpy_rng.uniform(low=[-4.], high=[-1])) numpy_val1b = as_floatX(numpy_rng.uniform(low=[-4.], high=[-1]))
assert numpy.all(val0b == numpy_val0b) assert np.all(val0b == numpy_val0b)
assert numpy.all(val1b == numpy_val1b) assert np.all(val1b == numpy_val1b)
self.assertRaises(ValueError, fb, post1b, [-4., -2], [-1, 0, 1]) self.assertRaises(ValueError, fb, post1b, [-4., -2], [-1, 0, 1])
# TODO: do we want that? # TODO: do we want that?
#self.assertRaises(ValueError, fb, post1b, [-4., -2], [-1]) #self.assertRaises(ValueError, fb, post1b, [-4., -2], [-1])
...@@ -681,8 +681,8 @@ class T_random_function(utt.InferShapeTester): ...@@ -681,8 +681,8 @@ class T_random_function(utt.InferShapeTester):
post1c, val1c = fc(post0c, [-4.], [-1], [1]) post1c, val1c = fc(post0c, [-4.], [-1], [1])
numpy_val0c = as_floatX(numpy_rng.uniform(low=[-4., -2], high=[-1, 0])) numpy_val0c = as_floatX(numpy_rng.uniform(low=[-4., -2], high=[-1, 0]))
numpy_val1c = as_floatX(numpy_rng.uniform(low=[-4.], high=[-1])) numpy_val1c = as_floatX(numpy_rng.uniform(low=[-4.], high=[-1]))
assert numpy.all(val0c == numpy_val0c) assert np.all(val0c == numpy_val0c)
assert numpy.all(val1c == numpy_val1c) assert np.all(val1c == numpy_val1c)
self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [1]) self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [1])
self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [1, 2]) self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [1, 2])
self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [2, 1]) self.assertRaises(ValueError, fc, post1c, [-4., -2], [-1, 0], [2, 1])
...@@ -699,8 +699,8 @@ class T_random_function(utt.InferShapeTester): ...@@ -699,8 +699,8 @@ class T_random_function(utt.InferShapeTester):
f = compile.function([rng_R, low, high], [post_r, out], f = compile.function([rng_R, low, high], [post_r, out],
accept_inplace=True) accept_inplace=True)
rng_state0 = numpy.random.RandomState(utt.fetch_seed()) rng_state0 = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
post0, val0 = f(rng_state0, [-5, .5, 0, 1], [[1.]]) post0, val0 = f(rng_state0, [-5, .5, 0, 1], [[1.]])
post1, val1 = f(post0, [.9], [[1.], [1.1], [1.5]]) post1, val1 = f(post0, [.9], [[1.], [1.1], [1.5]])
post2, val2 = f(post1, [-5, .5, 0, 1], [[1.], [1.1], [1.5]]) post2, val2 = f(post1, [-5, .5, 0, 1], [[1.], [1.1], [1.5]])
...@@ -710,9 +710,9 @@ class T_random_function(utt.InferShapeTester): ...@@ -710,9 +710,9 @@ class T_random_function(utt.InferShapeTester):
numpy_val2 = numpy_rng.uniform(low=[-5, .5, 0, 1], numpy_val2 = numpy_rng.uniform(low=[-5, .5, 0, 1],
high=[[1.], [1.1], [1.5]]) high=[[1.], [1.1], [1.5]])
assert numpy.all(val0 == numpy_val0), (val0, numpy_val0) assert np.all(val0 == numpy_val0), (val0, numpy_val0)
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
def test_uniform_vector(self): def test_uniform_vector(self):
rng_R = random_state_type() rng_R = random_state_type()
...@@ -724,22 +724,22 @@ class T_random_function(utt.InferShapeTester): ...@@ -724,22 +724,22 @@ class T_random_function(utt.InferShapeTester):
accept_inplace=True) accept_inplace=True)
def as_floatX(thing): def as_floatX(thing):
return numpy.asarray(thing, dtype=theano.config.floatX) return np.asarray(thing, dtype=theano.config.floatX)
low_val = as_floatX([.1, .2, .3]) low_val = as_floatX([.1, .2, .3])
high_val = as_floatX([1.1, 2.2, 3.3]) high_val = as_floatX([1.1, 2.2, 3.3])
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Arguments of size (3,) # Arguments of size (3,)
rng0, val0 = f(rng, low_val, high_val) rng0, val0 = f(rng, low_val, high_val)
numpy_val0 = as_floatX(numpy_rng.uniform(low=low_val, high=high_val)) numpy_val0 = as_floatX(numpy_rng.uniform(low=low_val, high=high_val))
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
# arguments of size (2,) # arguments of size (2,)
rng1, val1 = f(rng0, low_val[:-1], high_val[:-1]) rng1, val1 = f(rng0, low_val[:-1], high_val[:-1])
numpy_val1 = as_floatX(numpy_rng.uniform(low=low_val[:-1], numpy_val1 = as_floatX(numpy_rng.uniform(low=low_val[:-1],
high=high_val[:-1])) high=high_val[:-1]))
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
# Specifying the size explicitly # Specifying the size explicitly
g = compile.function([rng_R, low, high], g = compile.function([rng_R, low, high],
...@@ -748,7 +748,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -748,7 +748,7 @@ class T_random_function(utt.InferShapeTester):
rng2, val2 = g(rng1, low_val, high_val) rng2, val2 = g(rng1, low_val, high_val)
numpy_val2 = as_floatX(numpy_rng.uniform(low=low_val, high=high_val, numpy_val2 = as_floatX(numpy_rng.uniform(low=low_val, high=high_val,
size=(3,))) size=(3,)))
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
self.assertRaises(ValueError, g, rng2, low_val[:-1], high_val[:-1]) self.assertRaises(ValueError, g, rng2, low_val[:-1], high_val[:-1])
def test_binomial_vector(self): def test_binomial_vector(self):
...@@ -761,19 +761,19 @@ class T_random_function(utt.InferShapeTester): ...@@ -761,19 +761,19 @@ class T_random_function(utt.InferShapeTester):
accept_inplace=True) accept_inplace=True)
n_val = [1, 2, 3] n_val = [1, 2, 3]
prob_val = numpy.asarray([.1, .2, .3], dtype=config.floatX) prob_val = np.asarray([.1, .2, .3], dtype=config.floatX)
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Arguments of size (3,) # Arguments of size (3,)
rng0, val0 = f(rng, n_val, prob_val) rng0, val0 = f(rng, n_val, prob_val)
numpy_val0 = numpy_rng.binomial(n=n_val, p=prob_val) numpy_val0 = numpy_rng.binomial(n=n_val, p=prob_val)
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
# arguments of size (2,) # arguments of size (2,)
rng1, val1 = f(rng0, n_val[:-1], prob_val[:-1]) rng1, val1 = f(rng0, n_val[:-1], prob_val[:-1])
numpy_val1 = numpy_rng.binomial(n=n_val[:-1], p=prob_val[:-1]) numpy_val1 = numpy_rng.binomial(n=n_val[:-1], p=prob_val[:-1])
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
# Specifying the size explicitly # Specifying the size explicitly
g = compile.function([rng_R, n, prob], g = compile.function([rng_R, n, prob],
...@@ -781,7 +781,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -781,7 +781,7 @@ class T_random_function(utt.InferShapeTester):
accept_inplace=True) accept_inplace=True)
rng2, val2 = g(rng1, n_val, prob_val) rng2, val2 = g(rng1, n_val, prob_val)
numpy_val2 = numpy_rng.binomial(n=n_val, p=prob_val, size=(3,)) numpy_val2 = numpy_rng.binomial(n=n_val, p=prob_val, size=(3,))
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
self.assertRaises(ValueError, g, rng2, n_val[:-1], prob_val[:-1]) self.assertRaises(ValueError, g, rng2, n_val[:-1], prob_val[:-1])
def test_normal_vector(self): def test_normal_vector(self):
...@@ -794,35 +794,35 @@ class T_random_function(utt.InferShapeTester): ...@@ -794,35 +794,35 @@ class T_random_function(utt.InferShapeTester):
accept_inplace=True) accept_inplace=True)
def as_floatX(thing): def as_floatX(thing):
return numpy.asarray(thing, dtype=theano.config.floatX) return np.asarray(thing, dtype=theano.config.floatX)
avg_val = [1, 2, 3] avg_val = [1, 2, 3]
std_val = as_floatX([.1, .2, .3]) std_val = as_floatX([.1, .2, .3])
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Arguments of size (3,) # Arguments of size (3,)
rng0, val0 = f(rng, avg_val, std_val) rng0, val0 = f(rng, avg_val, std_val)
numpy_val0 = as_floatX(numpy_rng.normal(loc=as_floatX(avg_val), numpy_val0 = as_floatX(numpy_rng.normal(loc=as_floatX(avg_val),
scale=as_floatX(std_val))) scale=as_floatX(std_val)))
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
# arguments of size (2,) # arguments of size (2,)
rng1, val1 = f(rng0, avg_val[:-1], std_val[:-1]) rng1, val1 = f(rng0, avg_val[:-1], std_val[:-1])
numpy_val1 = numpy.asarray(numpy_rng.normal(loc=avg_val[:-1], numpy_val1 = np.asarray(numpy_rng.normal(loc=avg_val[:-1],
scale=std_val[:-1]), scale=std_val[:-1]),
dtype=theano.config.floatX) dtype=theano.config.floatX)
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
# Specifying the size explicitly # Specifying the size explicitly
g = compile.function([rng_R, avg, std], g = compile.function([rng_R, avg, std],
normal(rng_R, avg=avg, std=std, size=(3,)), normal(rng_R, avg=avg, std=std, size=(3,)),
accept_inplace=True) accept_inplace=True)
rng2, val2 = g(rng1, avg_val, std_val) rng2, val2 = g(rng1, avg_val, std_val)
numpy_val2 = numpy.asarray(numpy_rng.normal(loc=avg_val, scale=std_val, numpy_val2 = np.asarray(numpy_rng.normal(loc=avg_val, scale=std_val,
size=(3,)), size=(3,)),
dtype=theano.config.floatX) dtype=theano.config.floatX)
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
self.assertRaises(ValueError, g, rng2, avg_val[:-1], std_val[:-1]) self.assertRaises(ValueError, g, rng2, avg_val[:-1], std_val[:-1])
def test_random_integers_vector(self): def test_random_integers_vector(self):
...@@ -836,29 +836,29 @@ class T_random_function(utt.InferShapeTester): ...@@ -836,29 +836,29 @@ class T_random_function(utt.InferShapeTester):
low_val = [100, 200, 300] low_val = [100, 200, 300]
high_val = [110, 220, 330] high_val = [110, 220, 330]
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Arguments of size (3,) # Arguments of size (3,)
rng0, val0 = f(rng, low_val, high_val) rng0, val0 = f(rng, low_val, high_val)
numpy_val0 = numpy.asarray([numpy_rng.randint(low=lv, high=hv+1) numpy_val0 = np.asarray([numpy_rng.randint(low=lv, high=hv+1)
for lv, hv in zip(low_val, high_val)]) for lv, hv in zip(low_val, high_val)])
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
# arguments of size (2,) # arguments of size (2,)
rng1, val1 = f(rng0, low_val[:-1], high_val[:-1]) rng1, val1 = f(rng0, low_val[:-1], high_val[:-1])
numpy_val1 = numpy.asarray([numpy_rng.randint(low=lv, high=hv+1) numpy_val1 = np.asarray([numpy_rng.randint(low=lv, high=hv+1)
for lv, hv in zip(low_val[:-1], high_val[:-1])]) for lv, hv in zip(low_val[:-1], high_val[:-1])])
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
# Specifying the size explicitly # Specifying the size explicitly
g = compile.function([rng_R, low, high], g = compile.function([rng_R, low, high],
random_integers(rng_R, low=low, high=high, size=(3,)), random_integers(rng_R, low=low, high=high, size=(3,)),
accept_inplace=True) accept_inplace=True)
rng2, val2 = g(rng1, low_val, high_val) rng2, val2 = g(rng1, low_val, high_val)
numpy_val2 = numpy.asarray([numpy_rng.randint(low=lv, high=hv+1) numpy_val2 = np.asarray([numpy_rng.randint(low=lv, high=hv+1)
for lv, hv in zip(low_val, high_val)]) for lv, hv in zip(low_val, high_val)])
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
self.assertRaises(ValueError, g, rng2, low_val[:-1], high_val[:-1]) self.assertRaises(ValueError, g, rng2, low_val[:-1], high_val[:-1])
# Vectorized permutation don't make sense: the only parameter, n, # Vectorized permutation don't make sense: the only parameter, n,
...@@ -875,30 +875,30 @@ class T_random_function(utt.InferShapeTester): ...@@ -875,30 +875,30 @@ class T_random_function(utt.InferShapeTester):
n_val = [1, 2, 3] n_val = [1, 2, 3]
pvals_val = [[.1, .9], [.2, .8], [.3, .7]] pvals_val = [[.1, .9], [.2, .8], [.3, .7]]
pvals_val = numpy.asarray(pvals_val, dtype=config.floatX) pvals_val = np.asarray(pvals_val, dtype=config.floatX)
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
# Arguments of size (3,) # Arguments of size (3,)
rng0, val0 = f(rng, n_val, pvals_val) rng0, val0 = f(rng, n_val, pvals_val)
numpy_val0 = numpy.asarray([numpy_rng.multinomial(n=nv, pvals=pv) numpy_val0 = np.asarray([numpy_rng.multinomial(n=nv, pvals=pv)
for nv, pv in zip(n_val, pvals_val)]) for nv, pv in zip(n_val, pvals_val)])
assert numpy.all(val0 == numpy_val0) assert np.all(val0 == numpy_val0)
# arguments of size (2,) # arguments of size (2,)
rng1, val1 = f(rng0, n_val[:-1], pvals_val[:-1]) rng1, val1 = f(rng0, n_val[:-1], pvals_val[:-1])
numpy_val1 = numpy.asarray([numpy_rng.multinomial(n=nv, pvals=pv) numpy_val1 = np.asarray([numpy_rng.multinomial(n=nv, pvals=pv)
for nv, pv in zip(n_val[:-1], pvals_val[:-1])]) for nv, pv in zip(n_val[:-1], pvals_val[:-1])])
assert numpy.all(val1 == numpy_val1) assert np.all(val1 == numpy_val1)
# Specifying the size explicitly # Specifying the size explicitly
g = compile.function([rng_R, n, pvals], g = compile.function([rng_R, n, pvals],
multinomial(rng_R, n=n, pvals=pvals, size=(3,)), multinomial(rng_R, n=n, pvals=pvals, size=(3,)),
accept_inplace=True) accept_inplace=True)
rng2, val2 = g(rng1, n_val, pvals_val) rng2, val2 = g(rng1, n_val, pvals_val)
numpy_val2 = numpy.asarray([numpy_rng.multinomial(n=nv, pvals=pv) numpy_val2 = np.asarray([numpy_rng.multinomial(n=nv, pvals=pv)
for nv, pv in zip(n_val, pvals_val)]) for nv, pv in zip(n_val, pvals_val)])
assert numpy.all(val2 == numpy_val2) assert np.all(val2 == numpy_val2)
self.assertRaises(ValueError, g, rng2, n_val[:-1], pvals_val[:-1]) self.assertRaises(ValueError, g, rng2, n_val[:-1], pvals_val[:-1])
def test_multinomial_tensor3_a(self): def test_multinomial_tensor3_a(self):
...@@ -914,15 +914,15 @@ class T_random_function(utt.InferShapeTester): ...@@ -914,15 +914,15 @@ class T_random_function(utt.InferShapeTester):
f = compile.function([rng_R, pvals], [post_r, out], f = compile.function([rng_R, pvals], [post_r, out],
accept_inplace=True) accept_inplace=True)
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
pvals_val = numpy.asarray([[[.1, .9], [.2, .8], [.3, .7]]]) pvals_val = np.asarray([[[.1, .9], [.2, .8], [.3, .7]]])
assert pvals_val.shape == (1, 3, 2) assert pvals_val.shape == (1, 3, 2)
new_rng, draw = f(rng, pvals_val) new_rng, draw = f(rng, pvals_val)
assert draw.shape == (1, 3, 2) assert draw.shape == (1, 3, 2)
assert numpy.allclose(draw.sum(axis=2), 9) assert np.allclose(draw.sum(axis=2), 9)
def test_multinomial_tensor3_b(self): def test_multinomial_tensor3_b(self):
# Test the examples given in the multinomial documentation regarding # Test the examples given in the multinomial documentation regarding
...@@ -937,15 +937,15 @@ class T_random_function(utt.InferShapeTester): ...@@ -937,15 +937,15 @@ class T_random_function(utt.InferShapeTester):
f = compile.function([rng_R, pvals], [post_r, out], f = compile.function([rng_R, pvals], [post_r, out],
accept_inplace=True) accept_inplace=True)
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
numpy_rng = numpy.random.RandomState(utt.fetch_seed()) numpy_rng = np.random.RandomState(utt.fetch_seed())
pvals_val = numpy.asarray([[[.1, .9], [.2, .8], [.3, .7]]]) pvals_val = np.asarray([[[.1, .9], [.2, .8], [.3, .7]]])
assert pvals_val.shape == (1, 3, 2) assert pvals_val.shape == (1, 3, 2)
out_rng, draw = f(rng, pvals_val) out_rng, draw = f(rng, pvals_val)
assert draw.shape == (10, 1, 3, 2) assert draw.shape == (10, 1, 3, 2)
assert numpy.allclose(draw.sum(axis=3), 9) assert np.allclose(draw.sum(axis=3), 9)
def test_dtype(self): def test_dtype(self):
rng_R = random_state_type() rng_R = random_state_type()
...@@ -956,13 +956,13 @@ class T_random_function(utt.InferShapeTester): ...@@ -956,13 +956,13 @@ class T_random_function(utt.InferShapeTester):
assert out.dtype == 'int8' assert out.dtype == 'int8'
f = compile.function([rng_R, low, high], [post_r, out]) f = compile.function([rng_R, low, high], [post_r, out])
rng = numpy.random.RandomState(utt.fetch_seed()) rng = np.random.RandomState(utt.fetch_seed())
rng0, val0 = f(rng, 0, 9) rng0, val0 = f(rng, 0, 9)
assert val0.dtype == 'int8' assert val0.dtype == 'int8'
rng1, val1 = f(rng0, 255, 257) rng1, val1 = f(rng0, 255, 257)
assert val1.dtype == 'int8' assert val1.dtype == 'int8'
assert numpy.all(abs(val1) <= 1) assert np.all(abs(val1) <= 1)
def test_dtype_normal_uniform_687(self): def test_dtype_normal_uniform_687(self):
# Regression test for #687. # Regression test for #687.
...@@ -978,7 +978,7 @@ class T_random_function(utt.InferShapeTester): ...@@ -978,7 +978,7 @@ class T_random_function(utt.InferShapeTester):
def test_infer_shape(self): def test_infer_shape(self):
rng_R = random_state_type() rng_R = random_state_type()
rng_R_val = numpy.random.RandomState(utt.fetch_seed()) rng_R_val = np.random.RandomState(utt.fetch_seed())
# no shape specified, default args # no shape specified, default args
post_r, out = uniform(rng_R) post_r, out = uniform(rng_R)
......
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