先下载安装

git clone https://github.com/deepchem/deepchem.git # Clone deepchem source code from GitHubcd deepchemgpu 1 bash scripts/install_deepchem_conda.sh deepchem # If you want GPU supportsource activate deepchempython setup.py install # Manual installnosetests -a !slow -v deepchem --nologcapture # Run tests

测试能通过才怪

问题① tensorflow

没装tensorflow就:

pip install tensorflow-gpu 1.13.1

装完后如果import tensorflow 的时候报错 ImportError: libcublas.so.10.0: cannot open shared object file: No such file or directory

证明cuda不是10.0

1.13.1的tf只认10.0

不管以前的cuda是什么版本, 再装一个10.0

然后三个选择

1.设置软连接

sudo ln -s /usr/local/cuda-10.0 usr/local/cuda

以后用什么版本就连什么版本 

搞不定的话参考 https://www.jianshu.com/p/6a6fbce9073f

2.修改环境变量

# 先用vim打开.bashrc文件vim ~/.bashrc# 再将下面三句话粘贴进去 注意 假如你的安装目录不是cuda-10.0 则换成你的CUDA目录名称 export PATH $PATH:/usr/local/cuda-10.0/bin export LD_LIBRARY_PATH $LD_LIBRARY_PATH:/usr/local/cuda-10.0/lib64/ export LIBRARY_PATH $LIBRARY_PATH:/usr/local/cuda-10.0/lib64 

搞不定就参考https://zhuanlan.zhihu.com/p/35834028

3.修改虚拟环境下的环境变量

比如只在deepchem环境下用cuda10.0, 可以只修改环境下的变量

source activate deepchemcd $CONDA_PREFIXmkdir -p ./etc/conda/activate.dmkdir -p ./etc/conda/deactivate.dtouch ./etc/conda/activate.d/env_vars.shtouch ./etc/conda/deactivate.d/env_vars.sh

把下面内容写进 ./etc/conda/activate.d/env_vars.sh

#!/bin/shexport PATH $PATH:/usr/local/cuda-10.0/bin export LD_LIBRARY_PATH $LD_LIBRARY_PATH:/usr/local/cuda-10.0/lib64/ export LIBRARY_PATH $LIBRARY_PATH:/usr/local/cuda-10.0/lib64 

把下面内容写进 ./etc/conda/deactivate.d/env_vars.sh
 

#!/bin/shunset PATHunset LD_LIBRARY_PATHunset LIBRARY_PATH

搞不定参见conda官方文档

https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#saving-environment-variables

到这再import tensorflow试试,再不行可能是cudnn版本不对,试试安装cudnn7.3

问题② 没有joblib

pip install joblib

接下来再出现任何不能import的问题,请确认你用的python内核是3.5

现在应该可以用了