'utf-8' codec can't decode byte 0x88 in position 9: invalid start byte

1

qGgxATi66o8
qGgxATi66o81132×856 173 KB

I run this and get this error

UnicodeDecodeError Traceback (most recent call last)
in ()
----> 1 net.to(device)

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\nn\modules\module.py in to(self, *args, **kwargs)
377 return t.to(device, dtype if t.is_floating_point() else None, non_blocking)
378
→ 379 return self._apply(convert)
380
381 def register_backward_hook(self, hook):

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\nn\modules\module.py in _apply(self, fn)
183 def _apply(self, fn):
184 for module in self.children():
→ 185 module._apply(fn)
186
187 for param in self._parameters.values():

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\nn\modules\module.py in _apply(self, fn)
189 # Tensors stored in modules are graph leaves, and we don’t
190 # want to create copy nodes, so we have to unpack the data.
→ 191 param.data = fn(param.data)
192 if param._grad is not None:
193 param._grad.data = fn(param._grad.data)

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\nn\modules\module.py in convert(t)
375
376 def convert(t):
→ 377 return t.to(device, dtype if t.is_floating_point() else None, non_blocking)
378
379 return self._apply(convert)

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\cuda_init_.py in _lazy_init()
160 _check_driver()
161 torch._C._cuda_init()
→ 162 _cudart = _load_cudart()
163 _cudart.cudaGetErrorName.restype = ctypes.c_char_p
164 _cudart.cudaGetErrorString.restype = ctypes.c_char_p

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\cuda_init_.py in _load_cudart()
57 # First check the main program for CUDA symbols
58 if platform.system() == ‘Windows’:
—> 59 lib = find_cuda_windows_lib()
60 else:
61 lib = ctypes.cdll.LoadLibrary(None)

c:\users\илья\appdata\local\programs\python\python36\lib\site-packages\torch\cuda_init_.py in find_cuda_windows_lib()
30 proc = Popen([‘where’, ‘cudart64*.dll’], stdout=PIPE, stderr=PIPE)
31 out, err = proc.communicate()
—> 32 out = out.decode().strip()
33 if len(out) > 0:
34 if out.find(‘\r\n’) != -1:

UnicodeDecodeError: ‘utf-8’ codec can’t decode byte 0x88 in position 9: invalid start byte

windows 10
python 3.6.4
cuda 9.0

2

Hm, this looks more like a OS/Python issue to me rather than a Pytorch-specific problem. I guess it’s because you have non-ascii characters in your paths (c:\users\илья\).

Maybe try running the code as a .py script adding the following at the top of the script:

#!/usr/bin/env python
# -*- coding: utf-8 -*-

If possible for debugging, I would try to install PyTorch in a directory that has a path with ascii-characters only. If that doesn’t help, we know that there’s a different issue, and otherwise, this might be something @smth knows more about.

3

thank you)
I changed user name and it all worked.