Skip to content

P
pytensor
提交 a1e290b7 authored 作者: Frédéric Bastien's avatar Frédéric Bastien
浏览文件
操作

Merge pull request #3128 from harlouci/flake8_sparse

Flake8 sparse
上级 11a78c73 21887b7d
隐藏空白字符变更
内嵌 并排
正在显示 包含 111 行增加 和 112 行删除
theano/sparse/basic.py
浏览文件 @ a1e290b7
......@@ -16,7 +16,7 @@ from six.moves import xrange
import scipy.sparse
import theano
from theano import gof, tensor, compile, scalar, config
from theano import gof, tensor, scalar, config
from theano.gradient import DisconnectedType
from theano.sparse.utils import hash_from_sparse
import theano.tests.unittest_tools as utt
......@@ -28,10 +28,10 @@ sparse_formats = ['csc', 'csr']
""" Types of sparse matrices to use for testing """
_mtypes = [scipy.sparse.csc_matrix, scipy.sparse.csr_matrix]
#_mtypes = [sparse.csc_matrix, sparse.csr_matrix, sparse.dok_matrix,
# _mtypes = [sparse.csc_matrix, sparse.csr_matrix, sparse.dok_matrix,
# sparse.lil_matrix, sparse.coo_matrix]
#* new class ``dia_matrix`` : the sparse DIAgonal format
#* new class ``bsr_matrix`` : the Block CSR format
# * new class ``dia_matrix`` : the sparse DIAgonal format
# * new class ``bsr_matrix`` : the Block CSR format
_mtype_to_str = {scipy.sparse.csc_matrix: "csc",
scipy.sparse.csr_matrix: "csr"}
......@@ -153,7 +153,8 @@ def verify_grad_sparse(op, pt, structured=False, *args, **kwargs):
:return: None
"""
conv_none = lambda x: x
def conv_none(x):
return x
def conv_csr(ind, indptr, shp):
def f(spdata):
......@@ -358,9 +359,9 @@ class SparseVariable(_sparse_py_operators, gof.Variable):
def __str__(self):
return '%s{%s,%s}' % (
self.__class__.__name__,
self.format,
self.dtype)
self.__class__.__name__,
self.format,
self.dtype)
def __repr__(self):
return str(self)
......@@ -369,11 +370,11 @@ class SparseVariable(_sparse_py_operators, gof.Variable):
class SparseConstantSignature(tuple):
def __eq__(self, other):
(a, b), (x, y) = self, other
return a == x \
and (b.dtype == y.dtype)\
and (type(b) == type(y))\
and (b.shape == y.shape)\
and (abs(b - y).sum() < 1e-6 * b.nnz)
return (a == x and
(b.dtype == y.dtype) and
(type(b) == type(y)) and
(b.shape == y.shape) and
(abs(b - y).sum() < 1e-6 * b.nnz))
def __hash__(self):
(a, b) = self
......@@ -394,11 +395,11 @@ class SparseConstant(gof.Constant, _sparse_py_operators):
def __str__(self):
return '%s{%s,%s,shape=%s,nnz=%s}' % (
self.__class__.__name__,
self.format,
self.dtype,
self.data.shape,
self.data.nnz)
self.__class__.__name__,
self.format,
self.dtype,
self.data.shape,
self.data.nnz)
def __repr__(self):
return str(self)
......@@ -488,9 +489,10 @@ class CSMProperties(gof.Op):
csm = as_sparse_variable(csm)
assert csm.format in ["csr", "csc"]
data = tensor.TensorType(dtype=csm.type.dtype,
broadcastable=(False,)).make_variable()
broadcastable=(False,))()
return gof.Apply(self, [csm],
[data, tensor.ivector(), tensor.ivector(), tensor.ivector()])
[data, tensor.ivector(),
tensor.ivector(), tensor.ivector()])
def perform(self, node, inputs, out):
(csm,) = inputs
......@@ -646,7 +648,7 @@ class CSM(gof.Op):
return gof.Apply(self,
[data, indices, indptr, shape],
[SparseType(dtype=data.type.dtype,
format=self.format).make_variable()])
format=self.format)()])
def perform(self, node, inputs, outputs):
# for efficiency, if remap does nothing, then do not apply it
......@@ -658,7 +660,7 @@ class CSM(gof.Op):
if len(shape) != 2:
raise ValueError('Shape should be an array of length 2')
if (data.shape != indices.shape and numpy.size(data) !=
numpy.size(self.kmap)):
numpy.size(self.kmap)):
errmsg = ('Data (shape ' + repr(data.shape) +
' must have the same number of elements ' +
'as indices (shape' + repr(indices.shape) +
......@@ -684,7 +686,7 @@ class CSM(gof.Op):
g_data, g_indices, g_indptr, g_shape = csm_properties(g_out)
# unpack the data vector and wrap it as a 1d TensorType
g_data = csm_grad(self.kmap)(x_data, x_indices, x_indptr, x_shape,
g_data, g_indices, g_indptr, g_shape)
g_data, g_indices, g_indptr, g_shape)
return [g_data, DisconnectedType()(), DisconnectedType()(), DisconnectedType()()]
def infer_shape(self, node, shapes):
......@@ -776,14 +778,14 @@ class CSMGrad(gof.op.Op):
self.kmap)
def make_node(self, x_data, x_indices, x_indptr, x_shape,
g_data, g_indices, g_indptr, g_shape):
g_data, g_indices, g_indptr, g_shape):
gout_data = g_data.type()
return gof.Apply(self, [x_data, x_indices, x_indptr, x_shape,
g_data, g_indices, g_indptr, g_shape], [gout_data])
g_data, g_indices, g_indptr, g_shape], [gout_data])
def perform(self, node, inputs, outputs):
(x_data, x_indices, x_indptr, x_shape,
g_data, g_indices, g_indptr, g_shape) = inputs
g_data, g_indices, g_indptr, g_shape) = inputs
(g_out,) = outputs
if len(x_indptr) - 1 == x_shape[0]:
sp_dim = x_shape[1]
......@@ -834,7 +836,7 @@ class Cast(gof.op.Op):
assert x.format in ["csr", "csc"]
return gof.Apply(
self, [x],
[SparseType(dtype=self.out_type, format=x.format).make_variable()])
[SparseType(dtype=self.out_type, format=x.format)()])
def perform(self, node, inputs, outputs):
(x,) = inputs
......@@ -902,8 +904,8 @@ class DenseFromSparse(gof.op.Op):
self.sparse_grad = structured
def __eq__(self, other):
return (type(self) == type(other)) and \
(self.sparse_grad == other.sparse_grad)
return ((type(self) == type(other)) and
(self.sparse_grad == other.sparse_grad))
def __hash__(self):
return hash(type(self)) ^ hash(self.sparse_grad)
......@@ -918,8 +920,7 @@ class DenseFromSparse(gof.op.Op):
return gof.Apply(self,
[x],
[tensor.TensorType(dtype=x.type.dtype,
broadcastable=(False, False)
).make_variable()])
broadcastable=(False, False))()])
def perform(self, node, inputs, outputs):
(x,) = inputs
......@@ -990,8 +991,8 @@ class SparseFromDense(gof.op.Op):
x = tensor.as_tensor_variable(x)
if x.ndim > 2:
raise TypeError(
"Theano does not have sparse tensor types with more "
"than 2 dimensions, but %s.ndim = %i" % (x, x.ndim))
"Theano does not have sparse tensor types with more "
"than 2 dimensions, but %s.ndim = %i" % (x, x.ndim))
elif x.ndim == 1:
x = x.dimshuffle('x', 0)
elif x.ndim == 0:
......@@ -1002,8 +1003,7 @@ class SparseFromDense(gof.op.Op):
return gof.Apply(self,
[x],
[SparseType(dtype=x.type.dtype,
format=self.format
).make_variable()])
format=self.format)()])
def perform(self, node, inputs, outputs):
(x,) = inputs
......@@ -1243,7 +1243,7 @@ class GetItem2d(gof.op.Op):
# def infer_shape(self, node, i0_shapes):
# return i0_shapes
def make_node(self, x, index):
scipy_ver = [ int(n) for n in scipy.__version__.split('.')[:2]]
scipy_ver = [int(n) for n in scipy.__version__.split('.')[:2]]
x = as_sparse_variable(x)
assert x.format in ["csr", "csc"]
assert len(index) in [1, 2]
......@@ -1260,11 +1260,11 @@ class GetItem2d(gof.op.Op):
# If start or stop or step are None, make them a Generic
# constant. Else, they should be converted to Tensor Variables
# of dimension 1 and int/uint dtype.
if scipy_ver < [0, 14] and ind.step != None:
if scipy_ver < [0, 14] and ind.step is not None:
raise ValueError(
'Slice with step is not support with current'
' version of Scipy.')
if ind.step is None or ind.step == 1:
if ind.step is None or ind.step == 1:
step = generic_None
else:
if not isinstance(step, gof.Variable):
......@@ -1301,8 +1301,8 @@ class GetItem2d(gof.op.Op):
stop.ndim, stop.dtype)
elif ((isinstance(ind, gof.Variable) and
getattr(ind, 'ndim', -1) == 0)
or numpy.isscalar(ind)):
getattr(ind, 'ndim', -1) == 0) or
numpy.isscalar(ind)):
raise NotImplementedError(
'Theano has no sparse vector' +
'Use X[a:b, c:d], X[a:b, c:c+1] or X[a:b] instead.')
......@@ -1438,8 +1438,7 @@ class Transpose(gof.op.Op):
return gof.Apply(self,
[x],
[SparseType(dtype=x.type.dtype,
format=self.format_map[x.type.format]
).make_variable()])
format=self.format_map[x.type.format])()])
def perform(self, node, inputs, outputs):
(x,) = inputs
......@@ -1832,11 +1831,11 @@ class SquareDiagonal(gof.op.Op):
raise TypeError('data argument must be a vector', diag.type)
return gof.Apply(self, [diag],
[SparseType(dtype=diag.dtype, format='csc')()])
[SparseType(dtype=diag.dtype, format='csc')()])
def perform(self, node, inputs, outputs):
(z,) = outputs
diag, o_shape = inputs[0], inputs[0].shape * 2
diag = inputs[0]
N = len(diag)
data = diag[:N]
......@@ -1959,8 +1958,7 @@ class AddSS(gof.op.Op):
return gof.Apply(self,
[x, y],
[SparseType(dtype=out_dtype,
format=x.type.format
).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2001,7 +1999,7 @@ class AddSSData(gof.op.Op):
return gof.Apply(self,
[x, y],
[SparseType(dtype=x.type.dtype,
format=x.type.format).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2067,10 +2065,10 @@ class AddSD(gof.op.Op):
[x, y],
[tensor.TensorType(dtype=out_dtype,
broadcastable=y.type.broadcastable
).make_variable()])
)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
(x, y) = inputs
(out,) = outputs
assert _is_dense(y)
......@@ -2110,7 +2108,7 @@ class StructuredAddSV(gof.op.Op):
return gof.Apply(self,
[x, y],
[SparseType(dtype=x.type.dtype,
format=x.type.format).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2222,10 +2220,10 @@ class MulSS(gof.op.Op):
assert x.format in ["csr", "csc"]
assert y.format in ["csr", "csc"]
out_dtype = scalar.upcast(x.type.dtype, y.type.dtype)
return gof.Apply(self, [x, y],
return gof.Apply(self,
[x, y],
[SparseType(dtype=out_dtype,
format=x.type.format
)()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2300,7 +2298,6 @@ class MulSD(gof.op.Op):
out_dtype = node.outputs[0].dtype
if x.format == 'csc':
x_data = x.data
indices = x.indices
indptr = x.indptr
if x.dtype == out_dtype:
......@@ -2315,7 +2312,6 @@ class MulSD(gof.op.Op):
z_data[i_idx] *= y[i, j]
out[0] = z
elif x.format == 'csr':
x_data = x.data
indices = x.indices
indptr = x.indptr
if x.dtype == out_dtype:
......@@ -2363,12 +2359,12 @@ class MulSV(gof.op.Op):
if x.type.dtype != y.type.dtype:
raise NotImplementedError(
"MulSV not implemented for differing dtypes."
"Got %s and %s." % (str(x.type.dtype), str(y.type.dtype)))
"MulSV not implemented for differing dtypes."
"Got %s and %s." % (str(x.type.dtype), str(y.type.dtype)))
return gof.Apply(self,
[x, y],
[SparseType(dtype=x.type.dtype,
format=x.type.format).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2483,7 +2479,7 @@ class __ComparisonOpSS(gof.op.Op):
return gof.Apply(self,
[x, y],
[SparseType(dtype='uint8',
format=x.type.format).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2527,7 +2523,7 @@ class __ComparisonOpSD(gof.op.Op):
return gof.Apply(self,
[x, y],
[SparseType(dtype='uint8',
format=x.type.format).make_variable()])
format=x.type.format)()])
def perform(self, node, inputs, outputs):
(x, y) = inputs
......@@ -2765,9 +2761,10 @@ class HStack(gof.op.Op):
for x in var:
assert x.format in ["csr", "csc"]
return gof.Apply(
self, var,
[SparseType(dtype=self.dtype, format=self.format).make_variable()])
return gof.Apply(self,
var,
[SparseType(dtype=self.dtype,
format=self.format)()])
def perform(self, node, block, outputs):
(out,) = outputs
......@@ -2852,7 +2849,7 @@ class VStack(HStack):
def grad(self, inputs, gout):
(gz,) = gout
is_continuous = [(inputs[i].dtype in tensor.continuous_dtypes)
for i in range(len(inputs))]
for i in range(len(inputs))]
if _is_sparse_variable(gz):
gz = dense_from_sparse(gz)
......@@ -3213,8 +3210,7 @@ class TrueDot(gof.op.Op):
raise NotImplementedError()
inputs = [x, y] # Need to convert? e.g. assparse
outputs = [SparseType(dtype=x.type.dtype,
format=myformat).make_variable()]
outputs = [SparseType(dtype=x.type.dtype, format=myformat)()]
return gof.Apply(self, inputs, outputs)
def perform(self, node, inp, out_):
......@@ -3729,11 +3725,10 @@ def structured_dot_grad(sparse_A, dense_B, ga):
sdgcsx = sdg_csr
CSx = CSR
g_A_data = sdgcsx(csm_indices(sparse_A), \
g_A_data = sdgcsx(csm_indices(sparse_A),
csm_indptr(sparse_A), dense_B, ga)
return CSx(g_A_data, csm_indices(sparse_A), \
csm_indptr(sparse_A), \
csm_shape(sparse_A))
return CSx(g_A_data, csm_indices(sparse_A),
csm_indptr(sparse_A), csm_shape(sparse_A))
else:
raise NotImplementedError()
......@@ -3756,7 +3751,7 @@ class SamplingDot(gof.op.Op):
raise TypeError(p)
# TODO: use it.
dtype_out = scalar.upcast(x.type.dtype, y.type.dtype, p.type.dtype)
dtype_out = scalar.upcast(x.type.dtype, y.type.dtype, p.type.dtype) # noqa
return gof.Apply(self, [x, y, p], [p.type()])
......@@ -3857,7 +3852,8 @@ class Dot(gof.op.Op):
y_is_sparse_var = _is_sparse_variable(y)
if not x_is_sparse_var and not y_is_sparse_var:
raise TypeError("Sparse dot product should have at least one "
raise TypeError(
"Sparse dot product should have at least one "
"sparse variable as inputs, but the inputs are "
"%s (%s) and %s (%s)." % (x, x.type, y, y.type))
......
theano/sparse/opt.py
浏览文件 @ a1e290b7
......@@ -12,6 +12,7 @@ from theano.sparse import (CSC, CSR, csm_properties,
from theano.sparse import basic as sparse
_is_sparse_variable = sparse._is_sparse_variable
_is_dense = sparse._is_dense
# This is tested in tests/test_opt.py:test_local_csm_properties_csm
......@@ -47,10 +48,11 @@ def local_inplace_remove0(node):
return [new_node]
return False
theano.compile.optdb.register('local_inplace_remove0',
gof.TopoOptimizer(local_inplace_remove0,
failure_callback=gof.TopoOptimizer.warn_inplace),
60, 'fast_run', 'inplace')
theano.compile.optdb.register(
'local_inplace_remove0',
gof.TopoOptimizer(local_inplace_remove0,
failure_callback=gof.TopoOptimizer.warn_inplace),
60, 'fast_run', 'inplace')
class AddSD_ccode(gof.op.Op):
......@@ -174,10 +176,11 @@ def local_inplace_addsd_ccode(node):
inplace=True)(*node.inputs)
return [new_node]
return False
theano.compile.optdb.register('local_inplace_addsd_ccode',
gof.TopoOptimizer(local_inplace_addsd_ccode,
failure_callback=gof.TopoOptimizer.warn_inplace),
60, 'fast_run', 'inplace')
theano.compile.optdb.register(
'local_inplace_addsd_ccode',
gof.TopoOptimizer(local_inplace_addsd_ccode,
failure_callback=gof.TopoOptimizer.warn_inplace),
60, 'fast_run', 'inplace')
@register_canonicalize("fast_compile")
......@@ -234,16 +237,17 @@ class StructuredDotCSC(gof.Op):
def make_node(self, a_val, a_ind, a_ptr, a_nrows, b):
dtype_out = scalar.upcast(a_val.type.dtype, b.type.dtype)
r = gof.Apply(self, [a_val, a_ind, a_ptr, a_nrows, b],
[tensor.tensor(dtype_out, (False, b.type.broadcastable[1]))])
[tensor.tensor(dtype_out,
(False, b.type.broadcastable[1]))])
return r
def perform(self, node, inputs, outputs):
(a_val, a_ind, a_ptr, a_nrows, b) = inputs
(out,) = outputs
a = scipy.sparse.csc_matrix((a_val, a_ind, a_ptr),
(a_nrows, b.shape[0]),
copy=False)
#out[0] = a.dot(b)
(a_nrows, b.shape[0]),
copy=False)
# out[0] = a.dot(b)
out[0] = theano._asarray(a * b, dtype=node.outputs[0].type.dtype)
assert _is_dense(out[0]) # scipy 0.7 automatically converts to dense
......@@ -427,17 +431,18 @@ class StructuredDotCSR(gof.Op):
def make_node(self, a_val, a_ind, a_ptr, b):
self.dtype_out = scalar.upcast(a_val.type.dtype, b.type.dtype)
r = gof.Apply(self, [a_val, a_ind, a_ptr, b],
[tensor.tensor(self.dtype_out, (False,
b.type.broadcastable[1]))])
[tensor.tensor(self.dtype_out,
(False, b.type.broadcastable[1]))])
return r
def perform(self, node, inputs, outputs):
(a_val, a_ind, a_ptr, b) = inputs
(out,) = outputs
a = scipy.sparse.csr_matrix((a_val, a_ind, a_ptr),
(len(a_ptr) - 1, b.shape[0]),
copy=True) # use view_map before setting this to False
#out[0] = a.dot(b)
a = scipy.sparse.csr_matrix(
(a_val, a_ind, a_ptr),
(len(a_ptr) - 1, b.shape[0]),
copy=True) # use view_map before setting this to False
# out[0] = a.dot(b)
out[0] = a * b
# scipy 0.7 automatically converts to dense, but not .6 sometimes
assert _is_dense(out[0])
......@@ -634,7 +639,7 @@ class UsmmCscDense(gof.Op):
assert z.ndim == 2
dtype_out = scalar.upcast(alpha.type.dtype, x_val.type.dtype,
y.type.dtype, z.type.dtype)
y.type.dtype, z.type.dtype)
if dtype_out not in ('float32', 'float64'):
raise NotImplementedError('only float types are supported in '
......@@ -653,8 +658,9 @@ class UsmmCscDense(gof.Op):
if dtype_out != z.type.dtype:
z = tensor.cast(z, dtype_out)
r = gof.Apply(self, [alpha, x_val, x_ind, x_ptr, x_nrows, y, z],
[tensor.tensor(dtype_out, (False, y.type.broadcastable[1]))])
r = gof.Apply(
self, [alpha, x_val, x_ind, x_ptr, x_nrows, y, z],
[tensor.tensor(dtype_out, (False, y.type.broadcastable[1]))])
return r
def c_support_code(self):
......@@ -841,7 +847,7 @@ local_usmm = gof.opt.PatternSub(
{'pattern': 'alpha',
'constraint': lambda expr: (numpy.all(expr.type.broadcastable) and
theano.config.blas.ldflags)},
(sparse._dot, 'x', 'y'))),
(sparse._dot, 'x', 'y'))),
(usmm, (theano.tensor.neg, 'alpha'), 'x', 'y', 'z'))
register_specialize(local_usmm, name="local_usmm")
......@@ -896,7 +902,7 @@ class CSMGradC(gof.Op):
def make_node(self, a_val, a_ind, a_ptr, a_dim,
b_val, b_ind, b_ptr, b_dim):
return gof.Apply(self, [a_val, a_ind, a_ptr, a_dim,
b_val, b_ind, b_ptr, b_dim], [b_val.type()])
b_val, b_ind, b_ptr, b_dim], [b_val.type()])
def c_code(self, node, name, inputs, outputs, sub):
# retrieve dtype number
......@@ -1019,7 +1025,7 @@ def local_csm_grad_c(node):
return [csm_grad_c(*node.inputs)]
return False
# DISABLED AS IT IS BROKEN FOR UNSORTED INDICES!
#register_specialize(local_csm_grad_c, 'cxx_only')
# register_specialize(local_csm_grad_c, 'cxx_only')
class MulSDCSC(gof.Op):
......@@ -1572,7 +1578,7 @@ def local_structured_add_s_v(node):
x, y = node.inputs
x_is_sparse_variable = _is_sparse_variable(x)
#y_is_sparse_variable = _is_sparse_variable(y)
# y_is_sparse_variable = _is_sparse_variable(y)
if x_is_sparse_variable:
svar = x
......@@ -1840,7 +1846,7 @@ def local_sampling_dot_csr(node):
p_data, p_ind, p_ptr, p_shape = sparse.csm_properties(p)
z_data, z_ind, z_ptr = sampling_dot_csr(x, y, p_data,
p_ind, p_ptr, p_shape[1])
p_ind, p_ptr, p_shape[1])
return [sparse.CSR(z_data, z_ind, z_ptr, p_shape)]
return False
......
theano/sparse/type.py
浏览文件 @ a1e290b7
......@@ -100,8 +100,8 @@ class SparseType(gof.Type):
a, b = b, a
if _is_sparse(a) and isinstance(b, numpy.ndarray):
if (numpy.may_share_memory(a.data, b) or
numpy.may_share_memory(a.indices, b) or
numpy.may_share_memory(a.indptr, b)):
numpy.may_share_memory(a.indices, b) or
numpy.may_share_memory(a.indptr, b)):
# currently we can't share memory with a.shape as it is a tuple
return True
return False
......@@ -143,8 +143,8 @@ class SparseType(gof.Type):
# we definitely do not want to be doing this un-necessarily during
# a FAST_RUN computation..
return scipy.sparse.issparse(a) \
and scipy.sparse.issparse(b) \
and abs(a - b).sum() == 0.0
and scipy.sparse.issparse(b) \
and abs(a - b).sum() == 0.0
def is_valid_value(self, a):
return scipy.sparse.issparse(a) and (a.format == self.format)
......@@ -162,10 +162,10 @@ class SparseType(gof.Type):
# Register SparseType's C code for ViewOp.
theano.compile.register_view_op_c_code(
SparseType,
"""
Py_XDECREF(%(oname)s);
%(oname)s = %(iname)s;
Py_XINCREF(%(oname)s);
""",
1)
SparseType,
"""
Py_XDECREF(%(oname)s);
%(oname)s = %(iname)s;
Py_XINCREF(%(oname)s);
""",
1)
theano/tests/test_flake8.py
浏览文件 @ a1e290b7
......@@ -228,10 +228,7 @@ whitelist_flake8 = [
"misc/tests/test_pycuda_example.py",
"misc/hooks/reindent.py",
"misc/hooks/check_whitespace.py",
"sparse/type.py",
"sparse/__init__.py",
"sparse/opt.py",
"sparse/basic.py",
"sparse/tests/test_utils.py",
"sparse/tests/test_opt.py",
"sparse/tests/test_basic.py",
......
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