Fergal’s TF Ops¶
This library provides some convenience functions for doing some common operations in tensorflow. I recommend you also look at the tf.layers module, as there is a lot of overlap; this module aims to fill in the gaps that exist in tf.layers package.
If you are using tensorflow on a shared GPU server and want to control how many
GPUs it grabs, have a look py3nvml,
in particular the
Direct install from github (useful if you use pip freeze). To get the master branch, try:
$ pip install -e git+https://github.com/fbcotter/tf_ops#egg=tf_ops
or for a specific tag (e.g. 0.0.1), try:
$ pip install -e git+https://email@example.com#egg=tf_ops
Download and pip install from Git:
$ git clone https://github.com/fbcotter/tf_ops $ cd tf_ops $ pip install -r requirements.txt $ pip install -e .
I would recommend to download and install (with the editable flag), as it is likely you’ll want to tweak things/add functions more quickly than I can handle pull requests.