It would be better if this code didn't have the paths to different files hard-coded in. Maybe you could use the following commands to find the local path, and so locate the files you need:
import os print os.getcwd()

I suggest that you include the following details in your README file:

Install tensorflow (https://www.tensorflow.org/versions/r0.11/get_started/os_setup.html#download-and-setup).
Run GUI/main.py
Once open, type in a known redshift
Browse for any single spectrum FITS, ASCII, dat, or two-column text file.
Click any of the best matches to view the continuum-subtracted binned spectra.
If the input spectrum is too noisy, increase the smoothing level, and click 'Re-fit with priors'

A while ago Alex Kim suggested that the code should also be callable from the command line, returning a list of objects and probability. Is it possible to use the code in this way? So I would imagine the arguments would be the data file, and the redshift (if known), i.e.

python snclassifier.py DES16E2bkg_E2_combined_161005_v10_b00.dat 0.478

When I run without a known redshift, I get the following error

Traceback (most recent call last):
File "main.py", line 216, in list_item_clicked
self.plot_best_matches()
File "main.py", line 223, in plot_best_matches
wave_redshifted = self.wave * ((self.plotZ-self.knownZ) + 1)
AttributeError: 'MainApp' object has no attribute 'knownZ'

Maybe branch to a demo version, which reduced (but complete) functionality