我一直试图在keras中使用tf作为后端在centos 7服务器中运行NN，但是我收到了这个错误：
Using Theano backend.
WARNING (theano.configdefaults): install mkl with `conda install mkl-service`: No module named 'mkl'
Traceback (most recent call last):
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/configparser.py", line 1135, in _unify_values
sectiondict = self._sections[section]
KeyError: 'blas'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configparser.py", line 168, in fetch_val_for_key
return theano_cfg.get(section, option)
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/configparser.py", line 778, in get
d = self._unify_values(section, vars)
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/configparser.py", line 1138, in _unify_values
raise NoSectionError(section)
configparser.NoSectionError: No section: 'blas'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configparser.py", line 328, in __get__
delete_key=delete_key)
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configparser.py", line 172, in fetch_val_for_key
raise KeyError(key)
KeyError: 'blas.ldflags'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configdefaults.py", line 1250, in check_mkl_openmp
import mkl
ImportError: No module named 'mkl'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "ffnnecodata2.py", line 1, in <module>
from keras.layers import Input, Dense, LSTM, Activation, Dropout, Reshape
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/keras/__init__.py", line 3, in <module>
from . import utils
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/keras/utils/__init__.py", line 6, in <module>
from . import conv_utils
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/keras/utils/conv_utils.py", line 9, in <module>
from .. import backend as K
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/keras/backend/__init__.py", line 81, in <module>
from .theano_backend import *
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/keras/backend/theano_backend.py", line 7, in <module>
import theano
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/__init__.py", line 124, in <module>
from theano.scan_module import (scan, map, reduce, foldl, foldr, clone,
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/scan_module/__init__.py", line 41, in <module>
from theano.scan_module import scan_opt
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/scan_module/scan_opt.py", line 60, in <module>
from theano import tensor, scalar
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/tensor/__init__.py", line 17, in <module>
from theano.tensor import blas
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/tensor/blas.py", line 155, in <module>
from theano.tensor.blas_headers import blas_header_text
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/tensor/blas_headers.py", line 987, in <module>
if not config.blas.ldflags:
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configparser.py", line 332, in __get__
val_str = self.default()
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configdefaults.py", line 1447, in default_blas_ldflags
check_mkl_openmp()
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/theano/configdefaults.py", line 1262, in check_mkl_openmp
""")
RuntimeError:
Could not import 'mkl'. Either install mkl-service with conda or set
MKL_THREADING_LAYER=GNU in your environment for MKL 2018.
If you have MKL 2017 install and are not in a conda environment you
can set the Theano flag blas.check_openmp to False. Be warned that if
you set this flag and don't set the appropriate environment or make
sure you have the right version you *will* get wrong results.
我正在使用conda环境，我尝试安装mkl = 2017但它没有用。
我决定将我的代码翻译成tensorflow并尝试再次运行它，但我现在收到了这条消息：
Traceback (most recent call last):
File "tensorffnnecodata.py", line 17, in <module>
import tensorflow as tf
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/__init
.py", line 24, in <module>
from tensorflow.python import * # pylint: disable=redefined-builtin
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/python
_init__.py", line 49, in <module>
from tensorflow.python import pywrap_tensorflow
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/python
ywrap_tensorflow.py", line 74, in <module>
raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/python
ywrap_tensorflow.py", line 58, in <module>
from tensorflow.python.pywrap_tensorflow_internal import *
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/python
ywrap_tensorflow_internal.py", line 28, in <module>
_pywrap_tensorflow_internal = swig_import_helper()
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/site-packages/tensorflow/python
ywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/imp.py", line 243, in load_modu
return load_dynamic(name, filename, file)
File "/home/powerBuilding/anaconda3/envs/tensorgpu/lib/python3.5/imp.py", line 343, in load_dyna
c
return _load(spec)
ImportError: libcudnn.so.7: cannot open shared object file: No such file or directory
Failed to load the native TensorFlow runtime.
See https://www.tensorflow.org/install/install_sources#common_installation_problems
for some common reasons and solutions. Include the entire stack trace
above this error message when asking for help.
您知道我的安装有什么问题吗？
答案 0 :(得分：1)
该错误表明您需要为nVidia网站上安装的Cuda版本安装cuDNN 7.0（https://developer.nvidia.com/cudnn），并将文件复制到Cuda安装文件夹。 Tensorflow 1.7.0+的预构建二进制文件需要CUDA 9.0
ImportError: libcudnn.so.7: cannot open shared object file: No such file or directory