我尝试从python3虚拟环境中使用cuda构建tensorflow 1.10:
成功构建bazel的第一步(with the necessity to install additional things)
$ bazel build --config = opt --config = cuda // tensorflow / tools / pip_package:build_pip_package
第一次构建失败:
import keras_applications
ModuleNotFoundError: No module named 'keras_applications'
Target //tensorflow/tools/pip_package:build_pip_package failed to build
Use --verbose_failures to see the command lines of failed build steps.
INFO: Elapsed time: 9114.151s, Critical Path: 335.72s
INFO: 6696 processes: 6696 local.
FAILED: Build did NOT complete successfully
但是在安装了一些keras组件和h5py之后:
pip install keras_applications==1.0.4 --no-deps
Collecting keras_applications==1.0.4
Downloading https://files.pythonhosted.org/packages/54/90/8f327deaa37a71caddb59b7b4aaa9d4b3e90c0e76f8c2d1572005278ddc5/Keras_Applications-1.0.4-py2.py3-none-any.whl (43kB)
100% |████████████████████████████████| 51kB 633kB/s
Installing collected packages: keras-applications
Successfully installed keras-applications-1.0.4
(DeepFish) jeanpat@Dell-T5500:~/Developpement/Arch-TensorFlow/tensorflow$ pip install keras_preprocessing==1.0.2 --no-deps
Collecting keras_preprocessing==1.0.2
Downloading https://files.pythonhosted.org/packages/71/26/1e778ebd737032749824d5cba7dbd3b0cf9234b87ab5ec79f5f0403ca7e9/Keras_Preprocessing-1.0.2-py2.py3-none-any.whl
Installing collected packages: keras-preprocessing
Successfully installed keras-preprocessing-1.0.2
(DeepFish) jeanpat@Dell-T5500:~/Developpement/Arch-TensorFlow/tensorflow$ pip install h5py==2.8.0
Collecting h5py==2.8.0
Downloading https://files.pythonhosted.org/packages/8e/cb/726134109e7bd71d98d1fcc717ffe051767aac42ede0e7326fd1787e5d64/h5py-2.8.0-cp36-cp36m-manylinux1_x86_64.whl (2.8MB)
100% |████████████████████████████████| 2.8MB 1.9MB/s
Requirement already satisfied: numpy>=1.7 in /home/jeanpat/anaconda3/envs/DeepFish/lib/python3.6/site-packages (from h5py==2.8.0) (1.15.0)
Requirement already satisfied: six in /home/jeanpat/anaconda3/envs/DeepFish/lib/python3.6/site-packages (from h5py==2.8.0) (1.11.0)
keras-applications 1.0.4 has requirement keras>=2.1.6, but you'll have keras 2.0.8 which is incompatible.
Installing collected packages: h5py
Found existing installation: h5py 2.7.1
Uninstalling h5py-2.7.1:
Successfully uninstalled h5py-2.7.1
Successfully installed h5py-2.8.0
可以完成构建:
Target //tensorflow/tools/pip_package:build_pip_package up-to-date:
bazel-bin/tensorflow/tools/pip_package/build_pip_package
INFO: Elapsed time: 55.947s, Critical Path: 29.05s
INFO: 40 processes: 40 local.
INFO: Build completed successfully, 43 total actions
第二步构建也成功,但出现了一些警告:
$ bazel-bin / tensorflow / tools / pip_package / build_pip_package / tmp / tensorflow_pkg jeudi 23août2018,15:02:24(UTC + 0200):===在dir中准备源文件:/tmp/tmp.dEwVvSzrd0 〜/ Developpement / Arch-TensorFlow /张量流〜/ Developpement / Arch-TensorFlow /张量流 〜/开发/ Arch-TensorFlow / tensorflow jeudi 23août2018,15:02:31(UTC + 0200):===建筑车轮
警告:在目录“ ”
下找不到与“ .pd”匹配的文件警告:在目录“
” 下找不到与“ .dll”匹配的文件警告:在目录“
” 下找不到与“ .lib”匹配的文件警告:在目录“ tensorflow / include / tensorflow”下找不到与“ * .h”匹配的文件
警告:在目录“ tensorflow / include / Eigen”下找不到与“ *”匹配的文件
警告:在目录“ tensorflow / include / google”下找不到与“ * .h”匹配的文件
警告:在目录“ tensorflow / include / third_party”下找不到与“ *”匹配的文件
警告:在目录“ tensorflow / include / unsupported”下找不到与“ *”匹配的文件
jeudi 2018年8月23日,15:03:09(UTC + 0200):===输出车轮文件位于:/ tmp / tensorflow_pkg
最后,我尝试安装python软件包:
$ pip install /tmp/tensorflow_pkg/tensorflow-1.10.0-py2-none-any.whl
Requirement '/tmp/tensorflow_pkg/tensorflow-1.10.0-py2-none-any.whl' looks like a filename, but the file does not exist
tensorflow-1.10.0-py2-none-any.whl is not a supported wheel on this platform.
答案 0 :(得分:0)
软件包安装成功:
$ pip install /tmp/tensorflow_pkg/tensorflow*.whl