Create Conda Env
Create Conda Env
i) install anaconda, if you don't already have it:
Download Anaconda2 (suggested v5.2 for linux): https://www.continuum.io/downloads
$ cd Downloads
$ bash Anaconda2-5.2.0-Linux-x86_64.sh #file name might be slightly different, but follows this format
$ export PATH=~/anaconda2/bin:$PATH
or
$ source ./bashrc
software link: https://www.anaconda.com/distribution/#download-section
install doc: https://docs.anaconda.com/anaconda/install/linux/
Note that this install will modify the PATH variable in your bashrc.
You need to open a new terminal for that path change to take place (to be able to find conda
in the next step).
ii)create a conda env that will contain python 3:
$ conda create -n cs285_env python=3.5
(optional)delete env:
iii) activate the environment (do this every time you open a new terminal and want to run code):
$ source activate cs285_env
iv) install the requirements into this conda env
$ pip install --user --requirement requirements.txt
v) allow your code to be able to see 'cs285'
$ cd <path_to_hw>
$ pip install -e .
Note: This conda environment requires activating it every time you open a new terminal (in order to run code), but the benefit is that the required dependencies for this codebase will not affect existing/other versions of things on your computer. This stand-alone environment will have everything that is necessary.
vi) visualize saved tensorboard event file:
$ cd cs285/data/<your_log_dir>
$ tensorboard --logdir .
Then, navigate to shown url to see scalar summaries as plots (in 'scalar' tab), as well as videos (in 'images' tab)