Multi-layer Recurrent Neural Networks (LSTM, RNN) for character-level language models in Python using Tensorflow
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 94 commits ahead, 83 commits behind sherjilozair:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
data
php-keras-server
save-3d
save-abc
save-new4
save-tinyshakespeare
save
LICENSE.md
Procfile
README.md
black-off.png
black-on.png
heroku2.sh
index.php
model.py
requirements-backup.txt
requirements.txt
rnn-both-backup02.php
rnn-both-old-feb5-2017.php
rnn-both-resend.php
rnn-both.php
rnn-serve-test.sh
rnn-serve.html
rnn-serve.php
rnn-train-test.sh
rnn-train.php
sample-3d.py
sample-abc.py
sample-tinyshakespeare.1.py
sample-tinyshakespeare.py
sample.py
setup-oldfeb5-2017.sh
setup-test04.sh
setup-test05.sh
setup.sh
train.py
utils.py
white-off.png
white-on.png
x-train-3d3.py
x-train-abc.py

README.md

char-rnn-tensorflow

Multi-layer Recurrent Neural Networks (LSTM, RNN) for character-level language models in Python using Tensorflow.

Inspired from Andrej Karpathy's char-rnn.

.

.

.

.

#Updated By jeremy Ellis twitter @rocksetta Oct 24, 2016

Getting this working on cloud9 http://c9.io and with php support so that it can be done from a web page

Setup by making a blank workspace and running the setup.sh bash file.

Install this github

https://github.com/hpssjellis/char-rnn-tensorflow-music-3dprinting.git

I use a blank workspace but php5 or python would probably work.

To setup pythyon

right click --> run setup.py

To run the web server

Right click --> run the rnn-both.php file

(for a simple situtaion use rnn-serve.php with rnn-serve.html)

Open the link provided in the terminal output

I also have a github for installing both Magenta and this github onto cloud9 at

https://github.com/hpssjellis/my-tensorflow-magenta-online

use at your own risk! .

.

.

.

back to the original readme.md

.

Requirements

Basic Usage

To train with default parameters on the tinyshakespeare corpus, run python train.py.

To sample from a checkpointed model, python sample.py.

Roadmap

  • Add explanatory comments
  • Expose more command-line arguments
  • Compare accuracy and performance with char-rnn