Skip to content

P
pytensor
提交 1af6f15d authored 作者: Pascal Lamblin's avatar Pascal Lamblin 提交者： GitHub
浏览文件
操作

Merge pull request #5097 from slefrancois/test_gpuarray_absconv

use floatX in gpuarray dnn tests
上级 baa3dd12 b6747150
显示空白字符变更
内嵌 并排
正在显示 包含 154 行增加 和 128 行删除
theano/gpuarray/tests/test_dnn.py
浏览文件 @ 1af6f15d
...@@ -24,6 +24,22 @@ from .rnn_support import Model, GRU, LSTM, WrapperLayer ...@@ -24,6 +24,22 @@ from .rnn_support import Model, GRU, LSTM, WrapperLayer
from theano.configdefaults import SUPPORTED_DNN_CONV_ALGO_FWD from theano.configdefaults import SUPPORTED_DNN_CONV_ALGO_FWD
# If using float16, set CUDNN precision to float32
def set_precision(floatX):
if floatX == "float16":
precision = "float32"
else:
precision = theano.config.floatX
return precision
# If using float16, cast reference input to float32
def ref_cast(x):
if theano.config.floatX == 'float16':
x = T.cast(x, 'float32')
return x
def test_dnn_conv_desc_merge(): def test_dnn_conv_desc_merge():
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
...@@ -50,18 +66,17 @@ def test_dnn_conv_merge(): ...@@ -50,18 +66,17 @@ def test_dnn_conv_merge():
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img_shp = [2, 5, 6, 8] img_shp = [2, 5, 6, 8]
kern_shp = [3, 5, 5, 6] kern_shp = [3, 5, 5, 6]
img = T.ftensor4('img') img = T.tensor4('img')
kern = T.ftensor4('kern') kern = T.tensor4('kern')
out = T.ftensor4('out') out = T.tensor4('out')
desc = dnn.GpuDnnConvDesc( desc = dnn.GpuDnnConvDesc(border_mode='valid')(kern.shape)
border_mode='valid')(kern.shape)
# Test forward op # Test forward op
o1 = dnn.dnn_conv(img, kern) o1 = dnn.dnn_conv(img, kern)
o2 = dnn.dnn_conv(img, kern) o2 = dnn.dnn_conv(img, kern)
f = theano.function([img, kern], [o1, o2], mode=mode_with_gpu) f = theano.function([img, kern], [o1, o2], mode=mode_with_gpu)
d1, d2 = f(numpy.random.rand(*img_shp).astype('float32'), d1, d2 = f(numpy.random.rand(*img_shp).astype(theano.config.floatX),
numpy.random.rand(*kern_shp).astype('float32')) numpy.random.rand(*kern_shp).astype(theano.config.floatX))
topo = f.maker.fgraph.toposort() topo = f.maker.fgraph.toposort()
assert len([n for n in topo if isinstance(n.op, dnn.GpuDnnConv)]) == 1 assert len([n for n in topo if isinstance(n.op, dnn.GpuDnnConv)]) == 1
...@@ -89,9 +104,9 @@ def test_dnn_conv_inplace(): ...@@ -89,9 +104,9 @@ def test_dnn_conv_inplace():
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img_shp = [2, 5, 6, 8] img_shp = [2, 5, 6, 8]
kern_shp = [3, 5, 5, 6] kern_shp = [3, 5, 5, 6]
img = T.ftensor4('img') img = T.tensor4('img')
kern = T.ftensor4('kern') kern = T.tensor4('kern')
out = T.ftensor4('out') out = T.tensor4('out')
desc1 = dnn.GpuDnnConvDesc(border_mode='valid', conv_mode='conv')( desc1 = dnn.GpuDnnConvDesc(border_mode='valid', conv_mode='conv')(
kern.shape) kern.shape)
desc2 = dnn.GpuDnnConvDesc( desc2 = dnn.GpuDnnConvDesc(
...@@ -101,8 +116,8 @@ def test_dnn_conv_inplace(): ...@@ -101,8 +116,8 @@ def test_dnn_conv_inplace():
o1 = dnn.dnn_conv(img, kern, conv_mode='conv') o1 = dnn.dnn_conv(img, kern, conv_mode='conv')
o2 = dnn.dnn_conv(img, kern, conv_mode='cross') o2 = dnn.dnn_conv(img, kern, conv_mode='cross')
f = theano.function([img, kern], [o1, o2], mode=mode_with_gpu) f = theano.function([img, kern], [o1, o2], mode=mode_with_gpu)
d1, d2 = f(numpy.random.rand(*img_shp).astype('float32'), d1, d2 = f(numpy.random.rand(*img_shp).astype(theano.config.floatX),
numpy.random.rand(*kern_shp).astype('float32')) numpy.random.rand(*kern_shp).astype(theano.config.floatX))
topo = f.maker.fgraph.toposort() topo = f.maker.fgraph.toposort()
convs = [n for n in topo if isinstance(n.op, dnn.GpuDnnConv)] convs = [n for n in topo if isinstance(n.op, dnn.GpuDnnConv)]
assert len(convs) == 2 assert len(convs) == 2
...@@ -142,7 +157,7 @@ def test_pooling(): ...@@ -142,7 +157,7 @@ def test_pooling():
else: else:
modes = ('max', 'average_inc_pad', 'average_exc_pad') modes = ('max', 'average_inc_pad', 'average_exc_pad')
x = T.ftensor4() x = T.tensor4()
for mode, pad in product(modes, for mode, pad in product(modes,
((0, 0), (1, 0), (0, 1), (2, 3), (3, 2))): ((0, 0), (1, 0), (0, 1), (2, 3), (3, 2))):
if pad != (0, 0) and mode == 'average_exc_pad': if pad != (0, 0) and mode == 'average_exc_pad':
...@@ -180,7 +195,7 @@ def test_pooling(): ...@@ -180,7 +195,7 @@ def test_pooling():
(1, 3, 99, 99), (1, 3, 99, 99),
(32, 1, 147, 197), (32, 1, 147, 197),
]: ]:
data = numpy.random.normal(0, 1, shp).astype("float32") data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX)
a = f_cpu(data).__array__() a = f_cpu(data).__array__()
b = f_gpu(data).__array__() b = f_gpu(data).__array__()
utt.assert_allclose(a, b) utt.assert_allclose(a, b)
...@@ -188,7 +203,7 @@ def test_pooling(): ...@@ -188,7 +203,7 @@ def test_pooling():
# Test the grad # Test the grad
for shp in [(1, 1, 2, 2), for shp in [(1, 1, 2, 2),
(1, 1, 3, 3)]: (1, 1, 3, 3)]:
data = numpy.random.normal(0, 1, shp).astype("float32") * 10 data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX) * 10
ws = 2 ws = 2
stride = 2 stride = 2
...@@ -226,7 +241,7 @@ def test_pooling(): ...@@ -226,7 +241,7 @@ def test_pooling():
def test_pooling_with_tensor_vars(): def test_pooling_with_tensor_vars():
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
x = T.ftensor4() x = T.tensor4()
ws = theano.shared(numpy.array([2, 2], dtype='int32')) ws = theano.shared(numpy.array([2, 2], dtype='int32'))
stride = theano.shared(numpy.array([1, 1], dtype='int32')) stride = theano.shared(numpy.array([1, 1], dtype='int32'))
pad = theano.shared(numpy.array([0, 0], dtype='int32')) pad = theano.shared(numpy.array([0, 0], dtype='int32'))
...@@ -242,7 +257,7 @@ def test_pooling_with_tensor_vars(): ...@@ -242,7 +257,7 @@ def test_pooling_with_tensor_vars():
for shp in [(1, 1, 2, 2), for shp in [(1, 1, 2, 2),
(1, 1, 3, 3)]: (1, 1, 3, 3)]:
data = numpy.random.normal(0, 1, shp).astype("float32") * 10 data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX) * 10
theano.tests.unittest_tools.verify_grad( theano.tests.unittest_tools.verify_grad(
fn, [data], mode=mode_with_gpu) fn, [data], mode=mode_with_gpu)
...@@ -266,7 +281,7 @@ def test_pooling_with_tensor_vars(): ...@@ -266,7 +281,7 @@ def test_pooling_with_tensor_vars():
for shp in [(1, 10, 100, 100), for shp in [(1, 10, 100, 100),
(1, 3, 99, 99), (1, 3, 99, 99),
(32, 1, 147, 197)]: (32, 1, 147, 197)]:
data = numpy.random.normal(0, 1, shp).astype("float32") data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX)
# Change the window size dynamically # Change the window size dynamically
ws.set_value(numpy.array([i, i]).astype('int32')) ws.set_value(numpy.array([i, i]).astype('int32'))
...@@ -291,7 +306,7 @@ def test_pooling3d(): ...@@ -291,7 +306,7 @@ def test_pooling3d():
else: else:
modes = ('max', 'average_inc_pad', 'average_exc_pad') modes = ('max', 'average_inc_pad', 'average_exc_pad')
x = T.ftensor5() x = T.tensor5()
for mode, pad in product(modes, for mode, pad in product(modes,
((0, 0, 0), (1, 0, 0), (0, 1, 0), (0, 0, 1), ((0, 0, 0), (1, 0, 0), (0, 1, 0), (0, 0, 1),
(2, 3, 2), (3, 2, 2), (2, 2, 3))): (2, 3, 2), (3, 2, 2), (2, 2, 3))):
...@@ -327,11 +342,11 @@ def test_pooling3d(): ...@@ -327,11 +342,11 @@ def test_pooling3d():
(1, 3, 99, 99, 29), (1, 3, 99, 99, 29),
(2, 1, 147, 97, 37), (2, 1, 147, 97, 37),
]: ]:
data = numpy.random.normal(0, 1, shp).astype("float32") data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX)
a = f_cpu(data).__array__() a = f_cpu(data).__array__()
b = f_gpu(data).__array__() b = f_gpu(data).__array__()
utt.assert_allclose(a, b, utt.assert_allclose(a, b,
atol=numpy.finfo(numpy.float32).eps) atol=numpy.finfo(theano.config.floatX).eps)
# Test the grad # Test the grad
for shp in [(1, 1, 2, 2, 2), for shp in [(1, 1, 2, 2, 2),
...@@ -341,7 +356,7 @@ def test_pooling3d(): ...@@ -341,7 +356,7 @@ def test_pooling3d():
(1, 1, 4, 3, 3), (1, 1, 4, 3, 3),
(1, 1, 4, 4, 4), (1, 1, 4, 4, 4),
(1, 1, 5, 5, 5)]: (1, 1, 5, 5, 5)]:
data = numpy.random.normal(0, 1, shp).astype("float32") * 10 data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX) * 10
ws = 2 ws = 2
stride = 2 stride = 2
...@@ -370,7 +385,7 @@ def test_pooling_opt(): ...@@ -370,7 +385,7 @@ def test_pooling_opt():
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
# 2D pooling # 2D pooling
x = T.fmatrix() x = T.matrix()
f = theano.function( f = theano.function(
[x], [x],
...@@ -381,7 +396,7 @@ def test_pooling_opt(): ...@@ -381,7 +396,7 @@ def test_pooling_opt():
assert any([isinstance(n.op, dnn.GpuDnnPool) assert any([isinstance(n.op, dnn.GpuDnnPool)
for n in f.maker.fgraph.toposort()]) for n in f.maker.fgraph.toposort()])
f(numpy.zeros((10, 10), dtype='float32')) f(numpy.zeros((10, 10), dtype=theano.config.floatX))
# gradient of 2D pooling # gradient of 2D pooling
f = theano.function( f = theano.function(
...@@ -394,7 +409,7 @@ def test_pooling_opt(): ...@@ -394,7 +409,7 @@ def test_pooling_opt():
assert any([isinstance(n.op, dnn.GpuDnnPoolGrad) assert any([isinstance(n.op, dnn.GpuDnnPoolGrad)
for n in f.maker.fgraph.toposort()]) for n in f.maker.fgraph.toposort()])
f(numpy.zeros((10, 10), dtype='float32')) f(numpy.zeros((10, 10), dtype=theano.config.floatX))
# Test sum pooling # Test sum pooling
f = theano.function( f = theano.function(
...@@ -405,11 +420,11 @@ def test_pooling_opt(): ...@@ -405,11 +420,11 @@ def test_pooling_opt():
assert any([isinstance(n.op, dnn.GpuDnnPool) assert any([isinstance(n.op, dnn.GpuDnnPool)
for n in f.maker.fgraph.toposort()]) for n in f.maker.fgraph.toposort()])
data = numpy.random.rand(10, 10).astype('float32') data = numpy.random.rand(10, 10).astype(theano.config.floatX)
f(data) f(data)
# 3D pooling # 3D pooling
x = T.ftensor3() x = T.tensor3()
f = theano.function( f = theano.function(
[x], [x],
...@@ -420,7 +435,7 @@ def test_pooling_opt(): ...@@ -420,7 +435,7 @@ def test_pooling_opt():
assert any([isinstance(n.op, dnn.GpuDnnPool) assert any([isinstance(n.op, dnn.GpuDnnPool)
for n in f.maker.fgraph.toposort()]) for n in f.maker.fgraph.toposort()])
f(numpy.zeros((10, 10, 10), dtype='float32')) f(numpy.zeros((10, 10, 10), dtype=theano.config.floatX))
# gradient of 3D pooling # gradient of 3D pooling
f = theano.function( f = theano.function(
...@@ -433,7 +448,7 @@ def test_pooling_opt(): ...@@ -433,7 +448,7 @@ def test_pooling_opt():
assert any([isinstance(n.op, dnn.GpuDnnPoolGrad) assert any([isinstance(n.op, dnn.GpuDnnPoolGrad)
for n in f.maker.fgraph.toposort()]) for n in f.maker.fgraph.toposort()])
f(numpy.zeros((10, 10, 10), dtype='float32')) f(numpy.zeros((10, 10, 10), dtype=theano.config.floatX))
def test_pooling_opt_arbitrary_dimensions(): def test_pooling_opt_arbitrary_dimensions():
...@@ -454,7 +469,7 @@ def test_pooling_opt_arbitrary_dimensions(): ...@@ -454,7 +469,7 @@ def test_pooling_opt_arbitrary_dimensions():
# create input shape: non-pooling dimensions # create input shape: non-pooling dimensions
# followed by 2 or 3 pooling dimensions # followed by 2 or 3 pooling dimensions
shp = tuple(range(2, 2 + n_non_pool_dims)) + tuple(range(5, 5 + len(ws))) shp = tuple(range(2, 2 + n_non_pool_dims)) + tuple(range(5, 5 + len(ws)))
data = numpy.random.normal(0, 1, shp).astype('float32') data = numpy.random.normal(0, 1, shp).astype(theano.config.floatX)
input = gpuarray_shared_constructor(data) input = gpuarray_shared_constructor(data)
for mode in modes: for mode in modes:
...@@ -491,7 +506,7 @@ def test_dnn_tag(): ...@@ -491,7 +506,7 @@ def test_dnn_tag():
""" """
Test that if cudnn isn't avail we crash and that if it is avail, we use it. Test that if cudnn isn't avail we crash and that if it is avail, we use it.
""" """
x = T.ftensor4() x = T.tensor4()
old = theano.config.on_opt_error old = theano.config.on_opt_error
theano.config.on_opt_error = "raise" theano.config.on_opt_error = "raise"
...@@ -533,10 +548,10 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -533,10 +548,10 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_softmax(self): def test_softmax(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
t = T.ftensor4('t') t = T.tensor4('t')
rand_tensor = numpy.asarray( rand_tensor = numpy.asarray(
numpy.random.rand(5, 4, 3, 2), numpy.random.rand(5, 4, 3, 2),
dtype='float32' dtype=theano.config.floatX
) )
self._compile_and_check( self._compile_and_check(
[t], [t],
...@@ -564,19 +579,20 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -564,19 +579,20 @@ class TestDnnInferShapes(utt.InferShapeTester):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img_val = numpy.asarray(img_val, dtype='float32') img_val = numpy.asarray(img_val, dtype=theano.config.floatX)
kern_vals = numpy.asarray(kern_vals, dtype='float32') kern_vals = numpy.asarray(kern_vals, dtype=theano.config.floatX)
for subsample in subsamples: for subsample in subsamples:
out_vals = numpy.zeros( out_vals = numpy.zeros(
dnn.GpuDnnConv.get_out_shape(img_val.shape, kern_vals.shape, dnn.GpuDnnConv.get_out_shape(img_val.shape, kern_vals.shape,
border_mode=border_mode, border_mode=border_mode,
subsample=subsample), subsample=subsample),
dtype='float32') dtype=theano.config.floatX)
desc = dnn.GpuDnnConvDesc( desc = dnn.GpuDnnConvDesc(
border_mode=border_mode, border_mode=border_mode,
subsample=subsample, subsample=subsample,
conv_mode=conv_mode conv_mode=conv_mode,
precision=set_precision(theano.config.floatX)
)(kerns.shape) )(kerns.shape)
conv = dnn.GpuDnnConv(algo=algo)(img, kerns, out, desc) conv = dnn.GpuDnnConv(algo=algo)(img, kerns, out, desc)
self._compile_and_check( self._compile_and_check(
...@@ -597,9 +613,9 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -597,9 +613,9 @@ class TestDnnInferShapes(utt.InferShapeTester):
if algo == 'winograd' and dnn.version(raises=False) < 5000: if algo == 'winograd' and dnn.version(raises=False) < 5000:
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
self._test_conv(T.ftensor4('img'), self._test_conv(T.tensor4('img'),
T.ftensor4('kerns'), T.tensor4('kerns'),
T.ftensor4('out'), T.tensor4('out'),
numpy.random.rand(7, 2, 8, 4), numpy.random.rand(7, 2, 8, 4),
numpy.random.rand(8, 2, 4, 3), numpy.random.rand(8, 2, 4, 3),
border_mode, border_mode,
...@@ -609,9 +625,9 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -609,9 +625,9 @@ class TestDnnInferShapes(utt.InferShapeTester):
@parameterized.expand(product(border_modes, conv_modes), utt.custom_name_func) @parameterized.expand(product(border_modes, conv_modes), utt.custom_name_func)
def test_conv3d_none(self, border_mode, conv_mode): def test_conv3d_none(self, border_mode, conv_mode):
self._test_conv(T.ftensor5('img'), self._test_conv(T.tensor5('img'),
T.ftensor5('kerns'), T.tensor5('kerns'),
T.ftensor5('out'), T.tensor5('out'),
numpy.random.rand(10, 2, 6, 4, 11), numpy.random.rand(10, 2, 6, 4, 11),
numpy.random.rand(8, 2, 4, 3, 1), numpy.random.rand(8, 2, 4, 3, 1),
border_mode, border_mode,
...@@ -625,11 +641,11 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -625,11 +641,11 @@ class TestDnnInferShapes(utt.InferShapeTester):
img_val = numpy.asarray( img_val = numpy.asarray(
img_val, img_val,
dtype='float32' dtype=theano.config.floatX
) )
kern_vals = numpy.asarray( kern_vals = numpy.asarray(
kern_vals, kern_vals,
dtype='float32' dtype=theano.config.floatX
) )
temp_img = img.dimshuffle(1, 0, 2, 3) temp_img = img.dimshuffle(1, 0, 2, 3)
...@@ -642,11 +658,12 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -642,11 +658,12 @@ class TestDnnInferShapes(utt.InferShapeTester):
img_val.shape[2] - kern_vals.shape[2] + 1, img_val.shape[2] - kern_vals.shape[2] + 1,
img_val.shape[3] - kern_vals.shape[3] + 1 img_val.shape[3] - kern_vals.shape[3] + 1
) )
out_vals = numpy.zeros(shape, dtype='float32') out_vals = numpy.zeros(shape, dtype=theano.config.floatX)
desc = dnn.GpuDnnConvDesc( desc = dnn.GpuDnnConvDesc(
border_mode=border_mode, border_mode=border_mode,
subsample=subsample, subsample=subsample,
conv_mode=conv_mode conv_mode=conv_mode,
precision=set_precision(theano.config.floatX)
)(out.shape) )(out.shape)
conv_grad_w = dnn.GpuDnnConvGradW()( conv_grad_w = dnn.GpuDnnConvGradW()(
temp_img, temp_img,
...@@ -663,9 +680,9 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -663,9 +680,9 @@ class TestDnnInferShapes(utt.InferShapeTester):
@parameterized.expand(product(border_modes, conv_modes), utt.custom_name_func) @parameterized.expand(product(border_modes, conv_modes), utt.custom_name_func)
def test_conv_gradw(self, border_mode, conv_mode): def test_conv_gradw(self, border_mode, conv_mode):
self._test_conv_gradw(T.ftensor4('img'), self._test_conv_gradw(T.tensor4('img'),
T.ftensor4('kerns'), T.tensor4('kerns'),
T.ftensor4('out'), T.tensor4('out'),
numpy.random.rand(2, 5, 6, 8), numpy.random.rand(2, 5, 6, 8),
numpy.random.rand(2, 1, 5, 6), numpy.random.rand(2, 1, 5, 6),
border_mode, border_mode,
...@@ -675,16 +692,16 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -675,16 +692,16 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_conv_gradi(self): def test_conv_gradi(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor4('img') img = T.tensor4('img')
kerns = T.ftensor4('kerns') kerns = T.tensor4('kerns')
out = T.ftensor4('out') out = T.tensor4('out')
kern_vals = numpy.asarray( kern_vals = numpy.asarray(
numpy.random.rand(13, 14, 15, 16), numpy.random.rand(13, 14, 15, 16),
dtype='float32' dtype=theano.config.floatX
) )
out_vals = numpy.asarray( out_vals = numpy.asarray(
numpy.random.rand(3, 13, 5, 6), numpy.random.rand(3, 13, 5, 6),
dtype='float32' dtype=theano.config.floatX
) )
for params in product( for params in product(
...@@ -697,11 +714,12 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -697,11 +714,12 @@ class TestDnnInferShapes(utt.InferShapeTester):
out_vals.shape[2] + kern_vals.shape[2] - 1, out_vals.shape[2] + kern_vals.shape[2] - 1,
out_vals.shape[3] + kern_vals.shape[3] - 1 out_vals.shape[3] + kern_vals.shape[3] - 1
) )
img_vals = numpy.zeros(shape, dtype='float32') img_vals = numpy.zeros(shape, dtype=theano.config.floatX)
desc = dnn.GpuDnnConvDesc( desc = dnn.GpuDnnConvDesc(
border_mode=params[0], border_mode=params[0],
subsample=params[1], subsample=params[1],
conv_mode=params[2] conv_mode=params[2],
precision=set_precision(theano.config.floatX)
)(kerns.shape) )(kerns.shape)
conv_grad_i = dnn.GpuDnnConvGradI()( conv_grad_i = dnn.GpuDnnConvGradI()(
kerns, kerns,
...@@ -719,10 +737,10 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -719,10 +737,10 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_pool(self): def test_pool(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor4('img') img = T.tensor4('img')
img_val = numpy.asarray( img_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5), numpy.random.rand(2, 3, 4, 5),
dtype='float32' dtype=theano.config.floatX
) )
# 'average_exc_pad' is disabled for versions < 4004 # 'average_exc_pad' is disabled for versions < 4004
...@@ -746,10 +764,10 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -746,10 +764,10 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_pool_3d(self): def test_pool_3d(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor5('img') img = T.tensor5('img')
img_val = numpy.asarray( img_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5, 6), numpy.random.rand(2, 3, 4, 5, 6),
dtype='float32' dtype=theano.config.floatX
) )
# 'average_exc_pad' is disabled for versions < 4004 # 'average_exc_pad' is disabled for versions < 4004
...@@ -773,20 +791,20 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -773,20 +791,20 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_pool_grad(self): def test_pool_grad(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor4('img') img = T.tensor4('img')
img_grad = T.ftensor4('img_grad') img_grad = T.tensor4('img_grad')
out = T.ftensor4('out') out = T.tensor4('out')
img_val = numpy.asarray( img_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5), numpy.random.rand(2, 3, 4, 5),
dtype='float32' dtype=theano.config.floatX
) )
img_grad_val = numpy.asarray( img_grad_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5), numpy.random.rand(2, 3, 4, 5),
dtype='float32' dtype=theano.config.floatX
) )
out_val = numpy.asarray( out_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5), numpy.random.rand(2, 3, 4, 5),
dtype='float32' dtype=theano.config.floatX
) )
for params in product( for params in product(
...@@ -812,20 +830,20 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -812,20 +830,20 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_pool_3d_grad(self): def test_pool_3d_grad(self):
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor5('img') img = T.tensor5('img')
img_grad = T.ftensor5('img_grad') img_grad = T.tensor5('img_grad')
out = T.ftensor5('out') out = T.tensor5('out')
img_val = numpy.asarray( img_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5, 6), numpy.random.rand(2, 3, 4, 5, 6),
dtype='float32' dtype=theano.config.floatX
) )
img_grad_val = numpy.asarray( img_grad_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5, 6), numpy.random.rand(2, 3, 4, 5, 6),
dtype='float32' dtype=theano.config.floatX
) )
out_val = numpy.asarray( out_val = numpy.asarray(
numpy.random.rand(2, 3, 4, 5, 6), numpy.random.rand(2, 3, 4, 5, 6),
dtype='float32' dtype=theano.config.floatX
) )
for params in product( for params in product(
...@@ -853,8 +871,8 @@ class TestDnnInferShapes(utt.InferShapeTester): ...@@ -853,8 +871,8 @@ class TestDnnInferShapes(utt.InferShapeTester):
def test_dnn_conv_border_mode(): def test_dnn_conv_border_mode():
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor4() img = T.tensor4()
kern = T.ftensor4() kern = T.tensor4()
dnn.dnn_conv(img, kern, border_mode=1) dnn.dnn_conv(img, kern, border_mode=1)
dnn.dnn_conv(img, kern, border_mode=(2, 3)) dnn.dnn_conv(img, kern, border_mode=(2, 3))
...@@ -866,9 +884,9 @@ def test_dnn_conv_border_mode(): ...@@ -866,9 +884,9 @@ def test_dnn_conv_border_mode():
def test_dnn_conv_alpha_output_merge(): def test_dnn_conv_alpha_output_merge():
if not dnn.dnn_available(test_ctx_name): if not dnn.dnn_available(test_ctx_name):
raise SkipTest(dnn.dnn_available.msg) raise SkipTest(dnn.dnn_available.msg)
img = T.ftensor4() img = T.tensor4()
kern = T.ftensor4() kern = T.tensor4()
out = T.ftensor4() out = T.tensor4()
b = 1 b = 1
c = 4 c = 4
...@@ -877,16 +895,16 @@ def test_dnn_conv_alpha_output_merge(): ...@@ -877,16 +895,16 @@ def test_dnn_conv_alpha_output_merge():
iw = 8 iw = 8
kh = 2 kh = 2
kw = 6 kw = 6
img_val = numpy.random.random((b, c, ih, iw)).astype('float32') img_val = numpy.random.random((b, c, ih, iw)).astype(theano.config.floatX)
kern_val = numpy.random.random((f, c, kh, kw)).astype('float32') kern_val = numpy.random.random((f, c, kh, kw)).astype(theano.config.floatX)
out_val = numpy.random.random((b, f, ih - kh + 1, out_val = numpy.random.random((b, f, ih - kh + 1,
iw - kw + 1)).astype('float32') iw - kw + 1)).astype(theano.config.floatX)
conv = dnn.dnn_conv(img, kern) conv = dnn.dnn_conv(img, kern)
gw = theano.grad(conv.sum(), kern) gw = theano.grad(conv.sum(), kern)
gi = theano.grad(conv.sum(), img) gi = theano.grad(conv.sum(), img)
lr = numpy.asarray(0.05, dtype='float32') lr = numpy.asarray(0.05, dtype=theano.config.floatX)
fr = lr * (conv + out) fr = lr * (conv + out)
wr = kern + lr * gw wr = kern + lr * gw
...@@ -936,31 +954,31 @@ def test_dnn_conv_grad(): ...@@ -936,31 +954,31 @@ def test_dnn_conv_grad():
iw = 8 iw = 8
kh = 2 kh = 2
kw = 2 kw = 2
img_val = numpy.random.random((b, c, ih, iw)).astype('float32') img_val = numpy.random.random((b, c, ih, iw)).astype(theano.config.floatX)
kern_val = numpy.random.random((f, c, kh, kw)).astype('float32') kern_val = numpy.random.random((f, c, kh, kw)).astype(theano.config.floatX)
out_val = numpy.random.random((b, f, ih - kw + 1, out_val = numpy.random.random((b, f, ih - kw + 1,
iw - kw + 1)).astype('float32') iw - kw + 1)).astype(theano.config.floatX)
def dconv(img, kern, out): def dconv(img, kern, out):
desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1), desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1),
conv_mode='conv')(kern.shape) conv_mode='conv', precision=set_precision(theano.config.floatX))(kern.shape)
return dnn.GpuDnnConv()(img, kern, out, desc, alpha=0.5, beta=0.75) return dnn.GpuDnnConv()(img, kern, out, desc, alpha=0.5, beta=0.75)
def dconvi(img, kern, out): def dconvi(img, kern, out):
desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1), desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1),
conv_mode='conv')(kern.shape) conv_mode='conv', precision=set_precision(theano.config.floatX))(kern.shape)
return dnn.GpuDnnConvGradI()(kern, out, img, desc, alpha=-1.0, return dnn.GpuDnnConvGradI()(kern, out, img, desc, alpha=-1.0,
beta=0.0) beta=0.0)
def dconvw(img, kern, out): def dconvw(img, kern, out):
desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1), desc = dnn.GpuDnnConvDesc(border_mode='valid', subsample=(1, 1),
conv_mode='conv')(kern.shape) conv_mode='conv', precision=set_precision(theano.config.floatX))(kern.shape)
return dnn.GpuDnnConvGradW()(img, out, kern, desc, alpha=0.75, return dnn.GpuDnnConvGradW()(img, out, kern, desc, alpha=0.75,
beta=-1.0) beta=-1.0)
utt.verify_grad(dconv, [img_val, kern_val, out_val]) utt.verify_grad(dconv, [img_val, kern_val, out_val], eps=1e-3)
utt.verify_grad(dconvi, [img_val, kern_val, out_val]) utt.verify_grad(dconvi, [img_val, kern_val, out_val], eps=1e-3)
utt.verify_grad(dconvw, [img_val, kern_val, out_val]) utt.verify_grad(dconvw, [img_val, kern_val, out_val], eps=1e-3)
def get_conv3d_test_cases(): def get_conv3d_test_cases():
...@@ -1006,8 +1024,8 @@ def test_conv3d_fwd(): ...@@ -1006,8 +1024,8 @@ def test_conv3d_fwd():
def run_conv3d_fwd(inputs_shape, filters_shape, subsample, def run_conv3d_fwd(inputs_shape, filters_shape, subsample,
border_mode, conv_mode): border_mode, conv_mode):
inputs_val = numpy.random.random(inputs_shape).astype('float32') inputs_val = numpy.random.random(inputs_shape).astype(theano.config.floatX)
filters_val = numpy.random.random(filters_shape).astype('float32') filters_val = numpy.random.random(filters_shape).astype(theano.config.floatX)
# Scale down the input values to prevent very large absolute errors # Scale down the input values to prevent very large absolute errors
# due to float rounding # due to float rounding
...@@ -1033,13 +1051,18 @@ def test_conv3d_fwd(): ...@@ -1033,13 +1051,18 @@ def test_conv3d_fwd():
# Compile a theano function for the reference implementation # Compile a theano function for the reference implementation
conv_ref = theano.tensor.nnet.corr3d.Corr3dMM(border_mode=border_mode, conv_ref = theano.tensor.nnet.corr3d.Corr3dMM(border_mode=border_mode,
subsample=subsample subsample=subsample
)(inputs, flipped_filters) )(ref_cast(inputs), flipped_filters)
f_ref = theano.function([], conv_ref, mode="FAST_RUN") f_ref = theano.function([], conv_ref, mode="FAST_RUN")
# Compare the results of the two implementations # Compare the results of the two implementations
res_ref = f_ref() res_ref = f_ref()
res = f() res = f()
utt.assert_allclose(res_ref, res) # raise rtol to make the test pass with more seed.
rtol = None
# Raise tolerance for float16
if theano.config.floatX == 'float16':
rtol = 6e-2
utt.assert_allclose(res_ref, res, rtol=rtol)
test_cases = get_conv3d_test_cases() test_cases = get_conv3d_test_cases()
for (i_shape, f_shape, subsample), border_mode, conv_mode in test_cases: for (i_shape, f_shape, subsample), border_mode, conv_mode in test_cases:
...@@ -1055,8 +1078,8 @@ def test_conv3d_bwd(): ...@@ -1055,8 +1078,8 @@ def test_conv3d_bwd():
def run_conv3d_bwd(inputs_shape, filters_shape, subsample, def run_conv3d_bwd(inputs_shape, filters_shape, subsample,
border_mode, conv_mode): border_mode, conv_mode):
inputs_val = numpy.random.random(inputs_shape).astype('float32') inputs_val = numpy.random.random(inputs_shape).astype(theano.config.floatX)
filters_val = numpy.random.random(filters_shape).astype('float32') filters_val = numpy.random.random(filters_shape).astype(theano.config.floatX)
inputs = theano.shared(inputs_val) inputs = theano.shared(inputs_val)
filters = theano.shared(filters_val) filters = theano.shared(filters_val)
...@@ -1080,7 +1103,7 @@ def test_conv3d_bwd(): ...@@ -1080,7 +1103,7 @@ def test_conv3d_bwd():
# Compile a theano function for the reference implementation # Compile a theano function for the reference implementation
conv_ref = theano.tensor.nnet.corr3d.Corr3dMM(border_mode=border_mode, conv_ref = theano.tensor.nnet.corr3d.Corr3dMM(border_mode=border_mode,
subsample=subsample subsample=subsample
)(inputs, flipped_filters) )(ref_cast(inputs), flipped_filters)
(grad_i_ref, (grad_i_ref,
grad_w_ref) = theano.tensor.grad(conv_ref.sum(), grad_w_ref) = theano.tensor.grad(conv_ref.sum(),
[inputs, filters]) [inputs, filters])
...@@ -1091,8 +1114,12 @@ def test_conv3d_bwd(): ...@@ -1091,8 +1114,12 @@ def test_conv3d_bwd():
res = f() res = f()
# Needed for big size for some seed # Needed for big size for some seed
# raise rtol to make the test pass with more seed. # raise rtol to make the test pass with more seed.
utt.assert_allclose(res_ref[0], res[0], rtol=2e-5) rtol = None
utt.assert_allclose(res_ref[1], res[1], rtol=2e-5) # Raise tolerance for float16
if theano.config.floatX == 'float16':
rtol = 5e-2
utt.assert_allclose(res_ref[0], res[0], rtol=rtol)
utt.assert_allclose(res_ref[1], res[1], rtol=rtol)
test_cases = get_conv3d_test_cases() test_cases = get_conv3d_test_cases()
for (i_shape, f_shape, subsample), border_mode, conv_mode in test_cases: for (i_shape, f_shape, subsample), border_mode, conv_mode in test_cases:
...@@ -1139,15 +1166,15 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1139,15 +1166,15 @@ class test_SoftMax(test_nnet.test_SoftMax):
def test_softmax_grad(self): def test_softmax_grad(self):
def cmp(n, m, f, f_gpu): def cmp(n, m, f, f_gpu):
data = numpy.arange(n * m, dtype='float32').reshape(n, m) data = numpy.arange(n * m, dtype=theano.config.floatX).reshape(n, m)
gdata = numpy.asarray(data)[:, :, None, None] gdata = numpy.asarray(data)[:, :, None, None]
out = f(data) out = f(data)
gout = numpy.asarray(f_gpu(gdata))[:, :, 0, 0] gout = numpy.asarray(f_gpu(gdata))[:, :, 0, 0]
utt.assert_allclose(out, gout) utt.assert_allclose(out, gout)
x = T.matrix('x', 'float32') x = T.matrix('x')
x_gpu = T.tensor4('x_gpu', 'float32') x_gpu = T.tensor4('x_gpu')
f_z = T.nnet.softmax_op f_z = T.nnet.softmax_op
f_gpu = dnn.GpuDnnSoftmax( f_gpu = dnn.GpuDnnSoftmax(
'accurate', 'accurate',
...@@ -1158,7 +1185,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1158,7 +1185,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
dims = (2, 3, 4, 5) dims = (2, 3, 4, 5)
gdata = numpy.arange( gdata = numpy.arange(
numpy.product(dims), numpy.product(dims),
dtype='float32' dtype=theano.config.floatX
).reshape(dims) ).reshape(dims)
T.verify_grad(f_gpu, [gdata], rng=numpy.random, T.verify_grad(f_gpu, [gdata], rng=numpy.random,
mode=mode_with_gpu) mode=mode_with_gpu)
...@@ -1180,14 +1207,14 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1180,14 +1207,14 @@ class test_SoftMax(test_nnet.test_SoftMax):
# Verify that the SoftmaxGrad -> Gpu[Dnn]SoftmaxGrad # Verify that the SoftmaxGrad -> Gpu[Dnn]SoftmaxGrad
# optimization is applied when cudnn is required # optimization is applied when cudnn is required
y = T.fvector('y') y = T.vector('y')
f = theano.function( f = theano.function(
[y], [y],
T.grad(T.nnet.softmax(y).mean(), y), T.grad(T.nnet.softmax(y).mean(), y),
mode=mode_with_gpu mode=mode_with_gpu
) )
sorted_f = f.maker.fgraph.toposort() sorted_f = f.maker.fgraph.toposort()
val = numpy.random.rand(5).astype('float32') val = numpy.random.rand(5).astype(theano.config.floatX)
out_dnn = f(val) out_dnn = f(val)
assert(len([i assert(len([i
for i in sorted_f for i in sorted_f
...@@ -1206,7 +1233,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1206,7 +1233,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
# optimization is not applied when cudnn is excluded or not # optimization is not applied when cudnn is excluded or not
# available # available
mode_wo_cudnn = mode_with_gpu.excluding("cudnn") mode_wo_cudnn = mode_with_gpu.excluding("cudnn")
y = T.fvector('y') y = T.vector('y')
f = theano.function( f = theano.function(
[y], [y],
T.grad(T.nnet.softmax(y).mean(), y), T.grad(T.nnet.softmax(y).mean(), y),
...@@ -1230,7 +1257,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1230,7 +1257,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
# Verify that the SoftmaxGrad -> GpuDnnSoftmaxGrad do not # Verify that the SoftmaxGrad -> GpuDnnSoftmaxGrad do not
# crash with manual graph # crash with manual graph
y = T.fvector('y') y = T.vector('y')
o = theano.tensor.nnet.SoftmaxGrad()(y, y * 2) o = theano.tensor.nnet.SoftmaxGrad()(y, y * 2)
f = theano.function([y], o, mode=mode_with_gpu) f = theano.function([y], o, mode=mode_with_gpu)
sorted_f = f.maker.fgraph.toposort() sorted_f = f.maker.fgraph.toposort()
...@@ -1253,7 +1280,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1253,7 +1280,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
if dnn.version(raises=False) < 3000: if dnn.version(raises=False) < 3000:
raise SkipTest("Log-softmax is only in cudnn v3+") raise SkipTest("Log-softmax is only in cudnn v3+")
x = T.ftensor4() x = T.tensor4()
softmax_out = dnn.GpuDnnSoftmax('accurate', 'channel')(x) softmax_out = dnn.GpuDnnSoftmax('accurate', 'channel')(x)
log_out = T.log(T.as_tensor_variable(softmax_out)) log_out = T.log(T.as_tensor_variable(softmax_out))
...@@ -1277,7 +1304,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1277,7 +1304,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
(2, 3, 4, 66000)] (2, 3, 4, 66000)]
for inp_shape in input_shapes: for inp_shape in input_shapes:
input_val = numpy.random.normal(0, 1, inp_shape).astype("float32") input_val = numpy.random.normal(0, 1, inp_shape).astype(theano.config.floatX)
out = f(input_val) out = f(input_val)
expected_out = numpy.log(numpy.exp(input_val) / expected_out = numpy.log(numpy.exp(input_val) /
...@@ -1296,7 +1323,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1296,7 +1323,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
# Compile a reference function, on the CPU, to be used to validate the # Compile a reference function, on the CPU, to be used to validate the
# results of the other function. # results of the other function.
x = T.fmatrix() x = T.matrix()
f_ref = theano.function([x], T.nnet.LogSoftmax()(x)) f_ref = theano.function([x], T.nnet.LogSoftmax()(x))
# Build the first graph and ensure that the optimization is applied # Build the first graph and ensure that the optimization is applied
...@@ -1309,7 +1336,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1309,7 +1336,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
assert dnn_softmax_nodes[0].op.algo == "log" assert dnn_softmax_nodes[0].op.algo == "log"
# Compare the output of the function with the reference function # Compare the output of the function with the reference function
inp = numpy.random.normal(0, 1, (5, 6)).astype("float32") inp = numpy.random.normal(0, 1, (5, 6)).astype(theano.config.floatX)
utt.assert_allclose(f(inp), f_ref(inp)) utt.assert_allclose(f(inp), f_ref(inp))
# Build the first graph and ensure that the optimization is applied # Build the first graph and ensure that the optimization is applied
...@@ -1322,7 +1349,7 @@ class test_SoftMax(test_nnet.test_SoftMax): ...@@ -1322,7 +1349,7 @@ class test_SoftMax(test_nnet.test_SoftMax):
assert dnn_softmax_nodes[0].op.algo == "log" assert dnn_softmax_nodes[0].op.algo == "log"
# Compare the output of the function with the reference function # Compare the output of the function with the reference function
inp = numpy.random.normal(0, 1, (5, 6)).astype("float32") inp = numpy.random.normal(0, 1, (5, 6)).astype(theano.config.floatX)
utt.assert_allclose(f(inp), f_ref(inp)) utt.assert_allclose(f(inp), f_ref(inp))
...@@ -1334,7 +1361,7 @@ def test_dnn_batchnorm_train(): ...@@ -1334,7 +1361,7 @@ def test_dnn_batchnorm_train():
utt.seed_rng() utt.seed_rng()
for mode in ('per-activation', 'spatial'): for mode in ('per-activation', 'spatial'):
for vartype in (T.ftensor5, T.ftensor4, T.ftensor3, T.fmatrix, T.fvector): for vartype in (T.tensor5, T.tensor4, T.tensor3, T.matrix, T.vector):
x, scale, bias = (vartype(n) for n in ('x', 'scale', 'bias')) x, scale, bias = (vartype(n) for n in ('x', 'scale', 'bias'))
ndim = x.ndim ndim = x.ndim
eps = 5e-3 # some non-standard value to test if it's used eps = 5e-3 # some non-standard value to test if it's used
...@@ -1366,10 +1393,10 @@ def test_dnn_batchnorm_train(): ...@@ -1366,10 +1393,10 @@ def test_dnn_batchnorm_train():
data_shape = data_shape[:ndim] data_shape = data_shape[:ndim]
param_shape = tuple(1 if d in axes else s param_shape = tuple(1 if d in axes else s
for d, s in enumerate(data_shape)) for d, s in enumerate(data_shape))
X = 4 + 3 * numpy.random.randn(*data_shape).astype('float32') X = 4 + 3 * numpy.random.randn(*data_shape).astype(theano.config.floatX)
Dy = -1 + 2 * numpy.random.randn(*data_shape).astype('float32') Dy = -1 + 2 * numpy.random.randn(*data_shape).astype(theano.config.floatX)
Scale = numpy.random.randn(*param_shape).astype('float32') Scale = numpy.random.randn(*param_shape).astype(theano.config.floatX)
Bias = numpy.random.randn(*param_shape).astype('float32') Bias = numpy.random.randn(*param_shape).astype(theano.config.floatX)
outputs = f(X, Scale, Bias, Dy) outputs = f(X, Scale, Bias, Dy)
# compare outputs # compare outputs
utt.assert_allclose(outputs[0], outputs[0 + 3]) # out utt.assert_allclose(outputs[0], outputs[0 + 3]) # out
...@@ -1389,10 +1416,9 @@ def test_batchnorm_inference(): ...@@ -1389,10 +1416,9 @@ def test_batchnorm_inference():
utt.seed_rng() utt.seed_rng()
for mode in ('per-activation', 'spatial'): for mode in ('per-activation', 'spatial'):
for vartype in (T.ftensor5, T.ftensor4, T.ftensor3, T.fmatrix, T.fvector): for vartype in (T.tensor5, T.tensor4, T.tensor3, T.matrix, T.vector):
x, scale, bias, mean, var = (vartype(n) for n in ('x', 'scale', x, scale, bias, mean, var = (vartype(n)
'bias', 'mean', for n in ('x', 'scale', 'bias', 'mean', 'var'))
'var'))
ndim = x.ndim ndim = x.ndim
eps = 5e-3 # some non-standard value to test if it's used eps = 5e-3 # some non-standard value to test if it's used
...@@ -1420,12 +1446,12 @@ def test_batchnorm_inference(): ...@@ -1420,12 +1446,12 @@ def test_batchnorm_inference():
data_shape = data_shape[:ndim] data_shape = data_shape[:ndim]
param_shape = tuple(1 if d in axes else s param_shape = tuple(1 if d in axes else s
for d, s in enumerate(data_shape)) for d, s in enumerate(data_shape))
X = 4 + 3 * numpy.random.randn(*data_shape).astype('float32') X = 4 + 3 * numpy.random.randn(*data_shape).astype(theano.config.floatX)
Dy = -1 + 2 * numpy.random.randn(*data_shape).astype('float32') Dy = -1 + 2 * numpy.random.randn(*data_shape).astype(theano.config.floatX)
Scale = numpy.random.randn(*param_shape).astype('float32') Scale = numpy.random.randn(*param_shape).astype(theano.config.floatX)
Bias = numpy.random.randn(*param_shape).astype('float32') Bias = numpy.random.randn(*param_shape).astype(theano.config.floatX)
Mean = numpy.random.randn(*param_shape).astype('float32') Mean = numpy.random.randn(*param_shape).astype(theano.config.floatX)
Var = numpy.random.rand(*param_shape).astype('float32') Var = numpy.random.rand(*param_shape).astype(theano.config.floatX)
outputs = f(X, Scale, Bias, Mean, Var, Dy) outputs = f(X, Scale, Bias, Mean, Var, Dy)
# compare outputs # compare outputs
utt.assert_allclose(outputs[0], outputs[1]) # out utt.assert_allclose(outputs[0], outputs[1]) # out
......
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