ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

Question

ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

bokidigital opened this issue 5 years ago · comments

bokidigital commented 5 years ago

OS: ubuntu 18.04.1

After executed: /test.sh

I got:

ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

Brian Lee · Answer 1 · Tue Jun 18 2019 21:54:21 GMT+0800 (China Standard Time)

Have you installed CUDA 10?

bokidigital · Answer 2 · Wed Jun 19 2019 08:36:10 GMT+0800 (China Standard Time)

I installed it by this link:
https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=deblocal

Brian Lee · Answer 3 · Wed Jun 19 2019 22:05:18 GMT+0800 (China Standard Time)

You'll have to double check that all the .so objects are in the right place - the error message points at an incorrect or missing installation

bokidigital · Answer 4 · Thu Jun 20 2019 10:14:44 GMT+0800 (China Standard Time)

I did a clean installation, maybe README can provide a full of steps.

Tom Madams · Answer 5 · Thu Jun 20 2019 13:42:57 GMT+0800 (China Standard Time)

Sorry for the delayed response, I'm on vacation with limited internet access for the rest of the week. That link you posted is for CUDA 10.1, not 10.0.

The Minigo README.md does actually contain a link to the download page for the correct version of CUDA.

bokidigital · Answer 6 · Mon Jun 24 2019 15:09:04 GMT+0800 (China Standard Time)

I tested CUDA 10.0 in a clean system today; However, it got the same result as follows:
OS: ubuntu 18.04

./test.sh

///////////////////////////////////////////
During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "tests/run_tests.py", line 23, in
import test_dual_net
File "/home/rsa/minigo/minigo/tests/test_dual_net.py", line 20, in
import dual_net
File "./dual_net.py", line 29, in
import tensorflow as tf
File "/usr/local/lib/python3.6/dist-packages/tensorflow/init.py", line 24, in
from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/init.py", line 49, in
from tensorflow.python import pywrap_tensorflow
File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in
raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in
from tensorflow.python.pywrap_tensorflow_internal import *
File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in
_pywrap_tensorflow_internal = swig_import_helper()
File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
File "/usr/lib/python3.6/imp.py", line 243, in load_module
return load_dynamic(name, filename, file)
File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic
return _load(spec)
ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

Failed to load the native TensorFlow runtime.

See https://www.tensorflow.org/install/errors

Tom Madams · Answer 7 · Tue Jun 25 2019 01:20:10 GMT+0800 (China Standard Time)

What happens if you try importing TensorFlow directly instead of running the test script?

python -c "import tensorflow"

bokidigital · Answer 8 · Tue Jun 25 2019 09:41:43 GMT+0800 (China Standard Time)

python -c "import tensorflow"

Traceback (most recent call last):
File "", line 1, in
ImportError: No module named tensorflow

Tom Madams · Answer 9 · Tue Jun 25 2019 09:57:33 GMT+0800 (China Standard Time)

What about python3 -c "import tensorflow"

bokidigital · Answer 10 · Tue Jun 25 2019 11:02:20 GMT+0800 (China Standard Time)

python3 -c "import tensorflow" Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module> from tensorflow.python.pywrap_tensorflow_internal import * File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module> _pywrap_tensorflow_internal = swig_import_helper() File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description) File "/usr/lib/python3.6/imp.py", line 243, in load_module return load_dynamic(name, filename, file) File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic return _load(spec) ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory During handling of the above exception, another exception occurred: Traceback (most recent call last): File "<string>", line 1, in <module> File "/usr/local/lib/python3.6/dist-packages/tensorflow/__init__.py", line 24, in <module> from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/__init__.py", line 49, in <module> from tensorflow.python import pywrap_tensorflow File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 74, in <module> raise ImportError(msg) ImportError: Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module> from tensorflow.python.pywrap_tensorflow_internal import * File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module> _pywrap_tensorflow_internal = swig_import_helper() File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description) File "/usr/lib/python3.6/imp.py", line 243, in load_module return load_dynamic(name, filename, file) File "/usr/lib/python3.6/imp.py", line 343, in load_dynamic return _load(spec) ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory Failed to load the native TensorFlow runtime. See https://www.tensorflow.org/install/errors for some common reasons and solutions. Include the entire stack trace above this error message when asking for help.

…

On Tue, Jun 25, 2019 at 9:57 AM Tom Madams ***@***.***> wrote: What about python3 -c "import tensorflow" — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#843?email_source=notifications&email_token=ADGP23EQ4NGPRVQRI4JEIJ3P4F3RJA5CNFSM4HY5TPQKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYOXEKQ#issuecomment-505246250>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADGP23DKQTFF43BI5BIQRVDP4F3RJANCNFSM4HY5TPQA> .

Tom Madams · Answer 11 · Tue Jun 25 2019 11:30:15 GMT+0800 (China Standard Time)

You've got a problem with your TensorFlow install, which isn't really related to Minigo. I suggest you read through that doc if you haven't already done so and follow whatever its recommended steps are for getting help.

bokidigital · Answer 12 · Tue Jun 25 2019 13:48:00 GMT+0800 (China Standard Time)

I followed the steps in https://github.com/tensorflow/minigo to install tensorflow:

pip3 install "tensorflow-gpu==1.13.1"

on a clean system, actually, I don't know which step is wrong.

Tom Madams · Answer 13 · Wed Jun 26 2019 00:36:04 GMT+0800 (China Standard Time)

We don't maintain a full guide to installing tensorflow in our README, you should follow the guide on the tensorflow website: https://www.tensorflow.org/install/gpu

bokidigital · Answer 14 · Wed Jun 26 2019 09:13:31 GMT+0800 (China Standard Time)

After following the steps in:

https://www.tensorflow.org/install/gpu

The error mentioned in this thread is gone.

/////

2019-06-26 09:12:26.754448: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 15216 MB memory) -> physical GPU (device: 1, name: Tesla P100-PCIE-16GB, pci bus id: 0000:1e:00.0, compute capability: 6.0)
I0626 09:12:26.760015 139877677930304 dual_net.py:169] TPU inference is supported on C++ only. DualNetwork will ignore use_tpu=True
I0626 09:12:27.955729 139877677930304 saver.py:1270] Restoring parameters from /tmp/tmpmo7zaxss/models/000001-jupiter
I0626 09:12:28.051309 139877677930304 utils.py:86] Loading weights: 3.891 seconds
2019-06-26 09:12:28.090563: I tensorflow/stream_executor/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally
Finished game 0 W+1.5
Completed integration test!