m53297601
/
Paddle
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Fix unittests' division issues

Browse Source
revert-12469-sum_op_dim_fix
minqiyang 7 years ago
parent 59adf7ced1
commit be6ecec46f
5 changed files with 42 additions and 41 deletions
Show Stats Download Patch File Download Diff File

24
python/paddle/fluid/layers/nn.py
Unescape View File

@ -550,7 +550,7 @@ def dynamic_lstmp(input,
"""
helper = LayerHelper('lstmp', **locals())
size = size / 4
size = size // 4
weight = helper.create_parameter(
attr=helper.param_attr, shape=[proj_size, 4 * size], dtype=dtype)
proj_weight = helper.create_parameter(
@ -778,7 +778,7 @@ def gru_unit(input,
helper = LayerHelper('gru_unit', **locals())
dtype = helper.input_dtype()
size = size / 3
size = size // 3
# create weight
weight = helper.create_parameter(
@ -1258,7 +1258,7 @@ def sequence_conv(input,
outputs={"Out": pre_bias},
attrs={
'contextStride': filter_stride,
'contextStart': -int(filter_size / 2),
'contextStart': -int(filter_size // 2),
'contextLength': filter_size
})
pre_act = helper.append_bias_op(pre_bias)
@ -1487,7 +1487,7 @@ def conv2d(input,
else:
if num_channels % groups != 0:
raise ValueError("num_channels must be divisible by groups.")
num_filter_channels = num_channels / groups
num_filter_channels = num_channels // groups
filter_size = utils.convert_to_list(filter_size, 2, 'filter_size')
stride = utils.convert_to_list(stride, 2, 'stride')
@ -1649,7 +1649,7 @@ def conv3d(input,
else:
if num_channels % groups != 0:
raise ValueError("num_channels must be divisible by groups.")
num_filter_channels = num_channels / groups
num_filter_channels = num_channels // groups
filter_size = utils.convert_to_list(filter_size, 3, 'filter_size')
stride = utils.convert_to_list(stride, 3, 'stride')
@ -2384,16 +2384,16 @@ def conv2d_transpose(input,
w_in = input.shape[3]
filter_size_h = (output_size[0] - (h_in - 1) * stride[0] + 2 *
padding[0] - 1) / dilation[0] + 1
padding[0] - 1) // dilation[0] + 1
filter_size_w = (output_size[1] - (w_in - 1) * stride[1] + 2 *
padding[1] - 1) / dilation[1] + 1
padding[1] - 1) // dilation[1] + 1
filter_size = [filter_size_h, filter_size_w]
else:
filter_size = utils.convert_to_list(filter_size, 2,
'conv2d_transpose.filter_size')
groups = 1 if groups is None else groups
filter_shape = [input_channel, num_filters / groups] + filter_size
filter_shape = [input_channel, num_filters // groups] + filter_size
img_filter = helper.create_parameter(
dtype=input.dtype, shape=filter_shape, attr=helper.param_attr)
@ -2551,18 +2551,18 @@ def conv3d_transpose(input,
w_in = input.shape[4]
filter_size_d = (output_size[0] - (d_in - 1) * stride[0] + 2 *
padding[0] - 1) / dilation[0] + 1
padding[0] - 1) // dilation[0] + 1
filter_size_h = (output_size[1] - (h_in - 1) * stride[1] + 2 *
padding[1] - 1) / dilation[1] + 1
padding[1] - 1) // dilation[1] + 1
filter_size_w = (output_size[2] - (w_in - 1) * stride[2] + 2 *
padding[2] - 1) / dilation[2] + 1
padding[2] - 1) // dilation[2] + 1
filter_size = [filter_size_d, filter_size_h, filter_size_w]
else:
filter_size = utils.convert_to_list(filter_size, 3,
'conv3d_transpose.filter_size')
groups = 1 if groups is None else groups
filter_shape = [input_channel, num_filters / groups] + filter_size
filter_shape = [input_channel, num_filters // groups] + filter_size
img_filter = helper.create_parameter(
dtype=input.dtype, shape=filter_shape, attr=helper.param_attr)

18
python/paddle/fluid/tests/unittests/test_conv3d_op.py
Unescape View File

@ -24,14 +24,14 @@ def conv3d_forward_naive(input, filter, group, conv_param):
out_c, f_c, f_d, f_h, f_w = filter.shape
assert f_c * group == in_c
assert np.mod(out_c, group) == 0
sub_out_c = out_c / group
sub_out_c = out_c // group
stride, pad, dilation = conv_param['stride'], conv_param['pad'], conv_param[
'dilations']
out_d = 1 + (in_d + 2 * pad[0] - (dilation[0] * (f_d - 1) + 1)) / stride[0]
out_h = 1 + (in_h + 2 * pad[1] - (dilation[1] * (f_h - 1) + 1)) / stride[1]
out_w = 1 + (in_w + 2 * pad[2] - (dilation[2] * (f_w - 1) + 1)) / stride[2]
out_d = 1 + (in_d + 2 * pad[0] - (dilation[0] * (f_d - 1) + 1)) // stride[0]
out_h = 1 + (in_h + 2 * pad[1] - (dilation[1] * (f_h - 1) + 1)) // stride[1]
out_w = 1 + (in_w + 2 * pad[2] - (dilation[2] * (f_w - 1) + 1)) // stride[2]
out = np.zeros((in_n, out_c, out_d, out_h, out_w))
@ -166,7 +166,7 @@ class TestConv3dOp(OpTest):
self.stride = [1, 1, 1]
self.input_size = [2, 3, 4, 4, 4] # NCDHW
assert np.mod(self.input_size[1], self.groups) == 0
f_c = self.input_size[1] / self.groups
f_c = self.input_size[1] // self.groups
self.filter_size = [6, f_c, 3, 3, 3]
def init_dilation(self):
@ -185,7 +185,7 @@ class TestCase1(TestConv3dOp):
self.stride = [1, 1, 1]
self.input_size = [2, 3, 4, 4, 4] # NCDHW
assert np.mod(self.input_size[1], self.groups) == 0
f_c = self.input_size[1] / self.groups
f_c = self.input_size[1] // self.groups
self.filter_size = [6, f_c, 3, 3, 3]
@ -205,7 +205,7 @@ class TestWith1x1(TestConv3dOp):
self.stride = [1, 1, 1]
self.input_size = [2, 3, 4, 4, 4] # NCHW
assert np.mod(self.input_size[1], self.groups) == 0
f_c = self.input_size[1] / self.groups
f_c = self.input_size[1] // self.groups
self.filter_size = [6, f_c, 1, 1, 1]
def init_dilation(self):
@ -221,7 +221,7 @@ class TestWithInput1x1Filter1x1(TestConv3dOp):
self.stride = [1, 1, 1]
self.input_size = [2, 3, 1, 1, 1] # NCHW
assert np.mod(self.input_size[1], self.groups) == 0
f_c = self.input_size[1] / self.groups
f_c = self.input_size[1] // self.groups
self.filter_size = [6, f_c, 1, 1, 1]
def init_dilation(self):
@ -237,7 +237,7 @@ class TestWithDilation(TestConv3dOp):
self.stride = [1, 1, 1]
self.input_size = [2, 3, 6, 6, 6] # NCDHW
assert np.mod(self.input_size[1], self.groups) == 0
f_c = self.input_size[1] / self.groups
f_c = self.input_size[1] // self.groups
self.filter_size = [6, f_c, 2, 2, 2]
def init_dilation(self):

13
python/paddle/fluid/tests/unittests/test_infer_shape.py
Unescape View File

@ -14,6 +14,7 @@
import unittest
import six
import paddle.fluid.core as core
@ -27,14 +28,14 @@ class TestInferShape(unittest.TestCase):
shape = [10, 20]
# prepare input/output
x1 = block.var("x1")
x1 = block.var(six.b("x1"))
x1.set_type(core.VarDesc.VarType.LOD_TENSOR)
x1.set_shape(shape)
x2 = block.var("x2")
x2 = block.var(six.b("x2"))
x2.set_type(core.VarDesc.VarType.LOD_TENSOR)
x2.set_shape(shape)
out = block.var("out")
out = block.var(six.b("out"))
out.set_type(core.VarDesc.VarType.LOD_TENSOR)
# prepare the operator
@ -57,14 +58,14 @@ class TestInferShape(unittest.TestCase):
y_shape = [20, 30]
# prepare input/output
x1 = block.var("x")
x1 = block.var(six.b("x"))
x1.set_type(core.VarDesc.VarType.LOD_TENSOR)
x1.set_shape(x_shape)
x2 = block.var("y")
x2 = block.var(six.b("y"))
x2.set_type(core.VarDesc.VarType.LOD_TENSOR)
x2.set_shape(y_shape)
out = block.var("out")
out = block.var(six.b("out"))
out.set_type(core.VarDesc.VarType.LOD_TENSOR)
# prepare the operator

4
python/paddle/fluid/tests/unittests/test_layers.py
Unescape View File

@ -158,7 +158,7 @@ class TestBook(unittest.TestCase):
input=crf_decode,
label=label,
chunk_scheme="IOB",
num_chunk_types=(label_dict_len - 1) / 2)
num_chunk_types=(label_dict_len - 1) // 2)
self.assertFalse(crf is None)
self.assertFalse(crf_decode is None)
@ -285,7 +285,7 @@ class TestBook(unittest.TestCase):
name='word_{0}'.format(i), shape=[1], dtype='int64'))
dict_size = 10000
label_word = int(window_size / 2) + 1
label_word = int(window_size // 2) + 1
embs = []
for i in range(window_size):

24
python/paddle/fluid/tests/unittests/test_pool3d_op.py
Unescape View File

@ -29,14 +29,14 @@ def max_pool3D_forward_naive(x,
if global_pool == 1:
ksize = [D, H, W]
D_out = (D - ksize[0] + 2 * paddings[0] + strides[0] - 1
) / strides[0] + 1 if ceil_mode else (H - ksize[0] + 2 *
paddings[0]) / strides[0] + 1
) // strides[0] + 1 if ceil_mode else (H - ksize[0] + 2 *
paddings[0]) // strides[0] + 1
H_out = (H - ksize[1] + 2 * paddings[1] + strides[1] - 1
) / strides[1] + 1 if ceil_mode else (W - ksize[1] + 2 *
paddings[1]) / strides[1] + 1
) // strides[1] + 1 if ceil_mode else (W - ksize[1] + 2 *
paddings[1]) // strides[1] + 1
W_out = (W - ksize[2] + 2 * paddings[2] + strides[2] - 1
) / strides[2] + 1 if ceil_mode else (W - ksize[2] + 2 *
paddings[2]) / strides[2] + 1
) // strides[2] + 1 if ceil_mode else (W - ksize[2] + 2 *
paddings[2]) // strides[2] + 1
out = np.zeros((N, C, D_out, H_out, W_out))
for k in range(D_out):
d_start = np.max((k * strides[0] - paddings[0], 0))
@ -63,14 +63,14 @@ def avg_pool3D_forward_naive(x,
if global_pool == 1:
ksize = [D, H, W]
D_out = (D - ksize[0] + 2 * paddings[0] + strides[0] - 1
) / strides[0] + 1 if ceil_mode else (H - ksize[0] + 2 *
paddings[0]) / strides[0] + 1
) // strides[0] + 1 if ceil_mode else (H - ksize[0] + 2 *
paddings[0]) // strides[0] + 1
H_out = (H - ksize[1] + 2 * paddings[1] + strides[1] - 1
) / strides[1] + 1 if ceil_mode else (W - ksize[1] + 2 *
paddings[1]) / strides[1] + 1
) // strides[1] + 1 if ceil_mode else (W - ksize[1] + 2 *
paddings[1]) // strides[1] + 1
W_out = (W - ksize[2] + 2 * paddings[2] + strides[2] - 1
) / strides[2] + 1 if ceil_mode else (W - ksize[2] + 2 *
paddings[2]) / strides[2] + 1
) // strides[2] + 1 if ceil_mode else (W - ksize[2] + 2 *
paddings[2]) // strides[2] + 1
out = np.zeros((N, C, D_out, H_out, W_out))
for k in range(D_out):
d_start = np.max((k * strides[0] - paddings[0], 0))

Write Preview
Loading…
Cancel
Save