tensorlab / tensorfx

TensorFlow framework for training and serving machine learning models
Apache License 2.0
196 stars 41 forks source link

cannot train with ml engine/install it #11

Open brandondutra opened 7 years ago

brandondutra commented 7 years ago

Found this while trying cloud training. Error can be done locally with:

$ pwd .../tensorfx/build/dist $ pip install --upgrade --force-reinstall tensorfx-0.1.4.tar.gz Processing ./tensorfx-0.1.4.tar.gz Complete output from command python setup.py egg_info: Traceback (most recent call last): File "", line 1, in File "/tmp/pip-qs8P7n-build/setup.py", line 22, in with open('requirements.txt') as rf: IOError: [Errno 2] No such file or directory: 'requirements.txt'

----------------------------------------

Command "python setup.py egg_info" failed with error code 1 in /tmp/pip-qs8P7n-build/

brandondutra commented 7 years ago

problem: setuptools >= 34.3.1 is needed. I built the tar ball with 28.2.0 the first time and got this error.

brandondutra commented 7 years ago

sent PR #12 to check for this in build.sh. min setuptools needed is 30

nikhilk commented 7 years ago

Is this a pecularity of the internal goobuntu setup? I certainly didn't hit it on my mac where I use miniconda (but not a virtual or conda env).

brandondutra commented 7 years ago

which version of setuptools did you use?

nikhilk commented 7 years ago

I used whatever comes by standard with miniconda (as far as I remember, I didn't install it explicitly) which happens to be 34.2.0.

I checked on ubuntu. It is 26.1.1, and so this must be an ubuntu thing.

Does it make sense to require using miniconda as a starting point for the development enviornment, so there is a possibility of consistent environment?