Machine Learning Bash Setup

Hi All,

I am trying to set-up Non-GPU Machine learning Python setup. I was successfully able to install Tensorflow , Caffe,

But I am unable to setup Theano, OpenCV, Torch. They are all failing for admin access issues.

And I understand anything that has "Sudo" will not be installed.

One primary reason is to setup the Bash Console and share it with anyone who uses PythonAnywhere so that the resources and time is saved in going through the same setup. Why reinvent the wheel?

Excuse my grammar as its almost 2 AM while I am writing.

Any help or suggestions are welcome. Thanks to Google for routing me to PythonAnywhere as they are doing splendid job offering python environment for us. I wish Sudo is supported but thats too much to ask for free service.

Thanks PythonAnywhere. I atleast have few Machine Learning environment which I can access anywhere.

Regards Ssameerr

Hi there,

Thanks for the post!

For your installations, what exact command are you using to install them?

OpenCV is a tough one- because the standard installation actually sets up a new python executable with OpenCV integrations (ie. the python executable changes). This means it's pretty impossible to install as a user. We do however already have OpenCV installed for you (albeit perhaps an older version).

It seems like other people have been able to successfully install Torch on PythonAnywhere by using a virtualenv.

14:36 ~ $ pip install > /tmp/pytorch-install-log.txt HTTP error 403 while getting Could not install requirement torch==0.1.11.post5 from because of error 403 Client Error: Forbidden for url: Could not install requirement torch==0.1.11.post5 from because of HTTP error 403 Cli ent Error: Forbidden for url: for URL .11.post5-cp27-none-linux_x86_64.whl 17:33 ~ $ pip install torchvision Collecting torchvision Using cached torchvision-0.1.9-py2.py3-none-any.whl Requirement already satisfied (use --upgrade to upgrade): numpy in /usr/local/lib/python2.7/dist-packages (from torchvision) Requirement already satisfied (use --upgrade to upgrade): six in /usr/local/lib/python2.7/dist-packages (from torchvision) Collecting torch (from torchvision) Using cached torch-0.1.2.post1.tar.gz Complete output from command python egg_info: Traceback (most recent call last): File "<string>", line 1, in <module> File "/tmp/pip-build-xX0qGs/torch/", line 11, in <module> raise RuntimeError(README) RuntimeError: PyTorch does not currently provide packages for PyPI (see status at

Please follow the instructions at to install with miniconda instead.


Command "python egg_info" failed with error code 1 in /tmp/pip-build-xX0qGs/torch/ 17:33 ~ $ pip3.6 install > /tmp/pytorch-install-log.txt torch-0.1.11.post5-cp27-none-linux_x86_64.whl is not a supported wheel on this platform. 17:34 ~ $ pip install > /tmp/pytorch-install-log.txt

The prompt you pasted suggests that you didn't do that in a virtualenv. Do it in a virtualenv.