Dockerfile has broken numpy dependency due to pycuda/tensorflow incompatibility
dnns92 opened this issue · 3 comments
dnns92 commented
When using the provided dockerfile, pycuda and tensorflow cannot coexist, because numpy is incompatible. It must be < 1.20 due to the error message below for pycuda to work, but it must be >1.20 for tensorflow to work. Which numpy version did you use exactly?
See error log with numpy 1.24.2
python3 export.py --help
/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py:120: FutureWarning: In the future `np.bool` will be defined as the corresponding NumPy scalar.
reg.get_or_register_dtype("bool", np.bool)
Traceback (most recent call last):
File "export.py", line 9, in <module>
import pycuda.autoinit
File "/usr/local/lib/python3.8/dist-packages/pycuda/autoinit.py", line 7, in <module>
from pycuda.tools import make_default_context
File "/usr/local/lib/python3.8/dist-packages/pycuda/tools.py", line 49, in <module>
_fill_dtype_registry(respect_windows=True)
File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 221, in _fill_dtype_registry
fill_registry_with_c_types(
File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 120, in fill_registry_with_c_types
reg.get_or_register_dtype("bool", np.bool)
File "/usr/local/lib/python3.8/dist-packages/numpy/__init__.py", line 305, in __getattr__
raise AttributeError(__former_attrs__[attr])
AttributeError: module 'numpy' has no attribute 'bool'.
`np.bool` was a deprecated alias for the builtin `bool`. To avoid this error in existing code, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at:
https://numpy.org/devdocs/release/1.20.0-notes.html#deprecationsWhen using numpy 1.19
trtuser@85c39cba7ce6:/workspace$ python3 export.py
RuntimeError: module compiled against API version 0x10 but this version of numpy is 0xd . Check the section C-API incompatibility at the Troubleshooting ImportError section at https://numpy.org/devdocs/user/troubleshooting-importerror.html#c-api-incompatibility for indications on how to solve this problem .
Traceback (most recent call last):
File "export.py", line 8, in <module>
import pycuda.driver as cuda
File "/usr/local/lib/python3.8/dist-packages/pycuda/driver.py", line 62, in <module>
from pycuda._driver import * # noqa
SystemError: initialization of _driver raised unreported exceptionLishaTang commented
I also met this problem. Have u solved it?
Linaom1214 commented
When using the provided dockerfile, pycuda and tensorflow cannot coexist, because numpy is incompatible. It must be < 1.20 due to the error message below for pycuda to work, but it must be >1.20 for tensorflow to work. Which numpy version did you use exactly?
See error log with numpy 1.24.2
python3 export.py --help /usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py:120: FutureWarning: In the future `np.bool` will be defined as the corresponding NumPy scalar. reg.get_or_register_dtype("bool", np.bool) Traceback (most recent call last): File "export.py", line 9, in <module> import pycuda.autoinit File "/usr/local/lib/python3.8/dist-packages/pycuda/autoinit.py", line 7, in <module> from pycuda.tools import make_default_context File "/usr/local/lib/python3.8/dist-packages/pycuda/tools.py", line 49, in <module> _fill_dtype_registry(respect_windows=True) File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 221, in _fill_dtype_registry fill_registry_with_c_types( File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 120, in fill_registry_with_c_types reg.get_or_register_dtype("bool", np.bool) File "/usr/local/lib/python3.8/dist-packages/numpy/__init__.py", line 305, in __getattr__ raise AttributeError(__former_attrs__[attr]) AttributeError: module 'numpy' has no attribute 'bool'. `np.bool` was a deprecated alias for the builtin `bool`. To avoid this error in existing code, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here. The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecationsWhen using numpy 1.19
trtuser@85c39cba7ce6:/workspace$ python3 export.py RuntimeError: module compiled against API version 0x10 but this version of numpy is 0xd . Check the section C-API incompatibility at the Troubleshooting ImportError section at https://numpy.org/devdocs/user/troubleshooting-importerror.html#c-api-incompatibility for indications on how to solve this problem . Traceback (most recent call last): File "export.py", line 8, in <module> import pycuda.driver as cuda File "/usr/local/lib/python3.8/dist-packages/pycuda/driver.py", line 62, in <module> from pycuda._driver import * # noqa SystemError: initialization of _driver raised unreported exception
restart the container
Linaom1214 commented
When using the provided dockerfile, pycuda and tensorflow cannot coexist, because numpy is incompatible. It must be < 1.20 due to the error message below for pycuda to work, but it must be >1.20 for tensorflow to work. Which numpy version did you use exactly?
See error log with numpy 1.24.2python3 export.py --help /usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py:120: FutureWarning: In the future `np.bool` will be defined as the corresponding NumPy scalar. reg.get_or_register_dtype("bool", np.bool) Traceback (most recent call last): File "export.py", line 9, in <module> import pycuda.autoinit File "/usr/local/lib/python3.8/dist-packages/pycuda/autoinit.py", line 7, in <module> from pycuda.tools import make_default_context File "/usr/local/lib/python3.8/dist-packages/pycuda/tools.py", line 49, in <module> _fill_dtype_registry(respect_windows=True) File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 221, in _fill_dtype_registry fill_registry_with_c_types( File "/usr/local/lib/python3.8/dist-packages/pycuda/compyte/dtypes.py", line 120, in fill_registry_with_c_types reg.get_or_register_dtype("bool", np.bool) File "/usr/local/lib/python3.8/dist-packages/numpy/__init__.py", line 305, in __getattr__ raise AttributeError(__former_attrs__[attr]) AttributeError: module 'numpy' has no attribute 'bool'. `np.bool` was a deprecated alias for the builtin `bool`. To avoid this error in existing code, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here. The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecationsWhen using numpy 1.19
trtuser@85c39cba7ce6:/workspace$ python3 export.py RuntimeError: module compiled against API version 0x10 but this version of numpy is 0xd . Check the section C-API incompatibility at the Troubleshooting ImportError section at https://numpy.org/devdocs/user/troubleshooting-importerror.html#c-api-incompatibility for indications on how to solve this problem . Traceback (most recent call last): File "export.py", line 8, in <module> import pycuda.driver as cuda File "/usr/local/lib/python3.8/dist-packages/pycuda/driver.py", line 62, in <module> from pycuda._driver import * # noqa SystemError: initialization of _driver raised unreported exceptionrestart the container