ERROR: test_FitAllTypes (sknn.tests.test_types.TestMemoryMap) on win7 64

[gccrpm]

i installed scikit-neuralnetwork on windows7 64,and run :nosetests -v sknn.tests encounter:

ERROR: test_FitAllTypes (sknn.tests.test_types.TestMemoryMap)

Traceback (most recent call last):
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 99, in test_FitA
llTypes
X = self.make('X', (12, 3), dtype=t)
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 94, in make
return numpy.memmap(filename, dtype=dtype, mode='w+', shape=shape)
File "C:\Program Files\Enthought\Canopy\User\lib\site-packages\numpy\core\memm
ap.py", line 221, in new
fid = open(filename, (mode == 'c' and 'r' or mode)+'b')
IOError: [Errno 22] invalid mode ('w+b') or filename: 'c:\users\sks\appdata\
local\temp\tmptnxgh9\X'
-------------------- >> begin captured logging << --------------------
sknn: INFO: Initializing neural network with 1 layers, 3 inputs and 3 outputs.
sknn: DEBUG: - Dense: �[1;97mLinear �[0m Units: �[1;97m3 �[0m
sknn: DEBUG:
sknn: INFO: Training on dataset of 12 samples with 72 total size.
sknn: DEBUG: - Terminating loop after 1 total iterations.
1 0.000e+00 N/A 0.1s
sknn: DEBUG:
sknn: INFO: Terminating after specified 1 total iterations.
--------------------- >> end captured logging << ---------------------

======================================================================
ERROR: test_FitAllTypes (sknn.tests.test_types.TestMemoryMap)

Traceback (most recent call last):
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 90, in tearDown
shutil.rmtree(self.directory)
File "C:\Program Files\Enthought\Canopy\App\appdata\canopy-1.7.4.3348.win-x86_
64\lib\shutil.py", line 252, in rmtree
onerror(os.remove, fullname, sys.exc_info())
File "C:\Program Files\Enthought\Canopy\App\appdata\canopy-1.7.4.3348.win-x86_
64\lib\shutil.py", line 250, in rmtree
os.remove(fullname)
WindowsError: [Error 32] : 'c:\users\sks\appdata\local\temp\tmptnxgh9\X'
-------------------- >> begin captured logging << --------------------
sknn: INFO: Initializing neural network with 1 layers, 3 inputs and 3 outputs.
sknn: DEBUG: - Dense: �[1;97mLinear �[0m Units: �[1;97m3 �[0m
sknn: DEBUG:
sknn: INFO: Training on dataset of 12 samples with 72 total size.
sknn: DEBUG: - Terminating loop after 1 total iterations.
1 0.000e+00 N/A 0.1s
sknn: DEBUG:
sknn: INFO: Terminating after specified 1 total iterations.
--------------------- >> end captured logging << ---------------------

======================================================================
ERROR: test_PredictAllTypes (sknn.tests.test_types.TestMemoryMap)

Traceback (most recent call last):
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 106, in test_Pre
dictAllTypes
X = self.make('X', (12, 3), dtype=t)
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 94, in make
return numpy.memmap(filename, dtype=dtype, mode='w+', shape=shape)
File "C:\Program Files\Enthought\Canopy\User\lib\site-packages\numpy\core\memm
ap.py", line 221, in new
fid = open(filename, (mode == 'c' and 'r' or mode)+'b')
IOError: [Errno 22] invalid mode ('w+b') or filename: 'c:\users\sks\appdata\
local\temp\tmpabdgbn\X'
-------------------- >> begin captured logging << --------------------
sknn: WARNING: WARNING: Computing estimates with an untrained network.
sknn: INFO: Initializing neural network with 1 layers, 3 inputs and 3 outputs.
sknn: DEBUG: - Dense: �[1;97mLinear �[0m Units: �[1;97m3 �[0m
sknn: DEBUG:
--------------------- >> end captured logging << ---------------------

======================================================================
ERROR: test_PredictAllTypes (sknn.tests.test_types.TestMemoryMap)

Traceback (most recent call last):
File "D:\scikit-neuralnetwork\sknn\tests\test_types.py", line 90, in tearDown
shutil.rmtree(self.directory)
File "C:\Program Files\Enthought\Canopy\App\appdata\canopy-1.7.4.3348.win-x86_
64\lib\shutil.py", line 252, in rmtree
onerror(os.remove, fullname, sys.exc_info())
File "C:\Program Files\Enthought\Canopy\App\appdata\canopy-1.7.4.3348.win-x86_
64\lib\shutil.py", line 250, in rmtree
os.remove(fullname)
WindowsError: [Error 32] : 'c:\users\sks\appdata\local\temp\tmpabdgbn\X'
-------------------- >> begin captured logging << --------------------
sknn: WARNING: WARNING: Computing estimates with an untrained network.
sknn: INFO: Initializing neural network with 1 layers, 3 inputs and 3 outputs.
sknn: DEBUG: - Dense: �[1;97mLinear �[0m Units: �[1;97m3 �[0m
sknn: DEBUG:
--------------------- >> end captured logging << ---------------------

---

Ran 247 tests in 634.241s

FAILED (SKIP=10, errors=4)

it seems the numpy memmap go wrong, who can help me?