The models for the paper:

Hawkins J, Ahmad S (2016) Why Neurons Have Thousands of Synapses, a
Theory of Sequence Memory in Neocortex. Front Neural Circuits 10:23

are available here:

https://github.com/numenta/nupic

Notes provided by author Subutai Ahmad who also contributed to the
ModelDB notes below:

The simulations in the paper (specifically Figure 6) can be recreated
using the code here:
 
https://github.com/numenta/nupic.research/tree/master/projects/sequence_learning
 
In order to run it, the user would need to install our research
repository:
 
https://github.com/numenta/nupic.research
 
Which in turn depends on NuPIC:
 
https://github.com/numenta/nupic
 
We also have an active online forum for questions on the paper or the model:
 
https://discourse.numenta.org/c/htm-theory

---
Note from the ModeDB administrator:

I successfully installed the code on the unbuntu 14.04 platform.
In its default state it reproduced the red trace from Figure 6.

Since I was starting with a new ubuntu install, I installed the
following packages:

sudo apt-get install libfreetype6-dev libpng12-0 libpng12-dev libpng++-dev git python-setuptools python-devel python-numpy python-scipy build-essential gfortran autoconf automake libx11-dev

pip install https://s3-us-west-2.amazonaws.com/artifacts.numenta.org/numenta/nupic.core/releases/nupic.bindings/nupic.bindings-0.4.4-cp27-none-linux_x86_64.whl
pip install nupic

I found this forum helpful:

https://discourse.numenta.org/t/nupic-install-issue-on-ubuntu-16-04-lts/904

when I had to back track and start over. 

Once the nupic et al is successfully installed you can create a
results folder in the nupic.research/projects/sequence_learning folder

mkdir results

and then run with the command

python sequence_simulations.py

If you plot the columns "time" vs "accuracy", you will get the red
curve in Figure 6A. The other curves require different command line
options to that script.

I plotted the output in matlab after trimming off the header line with
the bash commands:

tail -8999 results/temp.csv > temp.dat
cat temp.dat | sed 's/,/ /g' > tmp.dat
cat tmp.dat | awk '{ print $1}' > t.dat
cat tmp.dat | awk '{ print $8}' > acc.dat

then in the matlab command prompt

load t.dat
load acc.dat
plot(t,acc,'r')

to produce the graph with a trace similar to the one in Figure 6