.. -- mode: rst --
|CircleCI| |ReadtheDocs|
.. |CircleCI| image:: https://circleci.com/gh/davidinouye/destructive-deep-learning/tree/master.svg?style=shield&circle-token=:circle-token .. _CircleCI: https://circleci.com/gh/davidinouye/destructive-deep-learning .. |ReadtheDocs| image:: https://readthedocs.org/projects/destructive-deep-learning/badge/?version=latest .. _ReadtheDocs: https://destructive-deep-learning.readthedocs.io/en/latest/?badge=latest
.. default-role:: any
Destructive deep learning estimators and functions. Estimators are compatible with scikit-learn. Source code is distributed under the BSD 3-clause license.
Please cite the following paper if you use this code:
| `Deep Density Destructors`_ (`pdf`_)
| David I. Inouye, Pradeep Ravikumar
| In *International Conference on Machine Learning* (ICML), 2018... Deep Density Destructors: http://www.cs.cmu.edu/~dinouye/papers/inouye2018-deep-density-destructors-icml2018.pdf
.. pdf: http://www.cs.cmu.edu/~dinouye/papers/inouye2018-deep-density-destructors-icml2018.pdf
UPDATED: Please see the updated documentation_ for an API reference and tutorials/demos including
a Quickstart tutorial and MNIST demo.
.. _updated documentation: https://destructive-deep-learning.readthedocs.io/en/latest/
Environment setup instructions for:
Docker or Singularity containers (recommended)
Linux Setup (Ubuntu)
Mac OSX (unsupported)
Docker or Singularity Setup
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Because MLPACK is required for the tree density destructors used in the experiments,
the suggested installation method is to download and start a shell in a Docker <https://www.docker.com/>
or Singularity <http://singularity.lbl.gov/> container as below.
(If you are using Docker for Mac or Docker for Windows_, you will probably have
to increase the available memory to Docker for these experiments. See Docker documentation.)
For Docker (recommended if available):
.. _MLPACK: http://mlpack.org/
.. _Docker for Mac: https://docs.docker.com/docker-for-mac/
.. _Docker for Windows: https://docs.docker.com/docker-for-windows/
.. code:: console
docker run -it davidinouye/destructive-deep-learning /bin/bashOr, for Singularity:
.. code:: console
singularity shell -s /bin/bash shub://davidinouye/destructive-deep-learning.. code:: console
apt-get update && apt-get install \
build-essential \
cmake \
liblapack-dev \
libblas-dev \
libboost-math-dev \
libboost-program-options-dev \
libboost-test-dev \
libboost-serialization-dev \
libarmadillo-devhomebrew as per homebrew documentation <https://brew.sh/>_ (note that this also install the required xcode tools).
Then, install cmake, armadillo and boost (required to build mlpack) and llvm (required for openmp support for mlpack):.. code:: console
brew update && brew install \
cmake \
armadillo \
boost \
llvmOnce your environment is setup via one of the methods described above, download and compile the code to link to MLPACK_.
The first pip install is for scikit-learn, cython is required to compile pot and ddl, and pot and nose are used in ddl tests.
.. code:: console
pip install numpy scipy scikit-learn
pip install setuptools Cython
git clone https://github.com/davidinouye/destructive-deep-learning.git
cd destructive-deep-learning
pip install .[test]To run tests (which uses pytest <https://docs.pytest.org/en/latest/>_), execute:
.. code:: console
make testNOTE: MLPACK_ is required to reproduce experiments, please
see installation instructions.
To reproduce the 2D experiment in the paper and generate the paper figures
open and run the notebook notebooks/demo_toy_experiment.ipynb <notebooks/demo_toy_experiment.ipynb>_
or run the notebook from the command line.
Note that this notebook may take a while to run.
Also, if the command below is interrupted with Ctrl+C, the underlying python process
may need to be killed manually.
.. code:: console
jupyter nbconvert --ExecutePreprocessor.timeout=-1 --to notebook --execute notebooks/demo_toy_experiment.ipynbTo reproduce the MNIST and CIFAR-10 experiments execute the command below.
Note that this script will download the MNIST and CIFAR-10 datasets into
data/download_cache if not downloaded already.
The results are stored in data/results both the log files and pickle files
that include the fitted models.
Note that the log files will always append to the previous log file rather
than overwriting the existing log file.
.. code:: console
# Download data cache directly since mldata.org is sometimes down
wget http://www.cs.cmu.edu/~dinouye/data/data-icml2018.tar.gz && tar -xzvf data-icml2018.tar.gz && rm data-icml2018.tar.gz
# Example command for deep copula model and MNIST data
python scripts/icml_2018_experiment.py --model_names=deep-copula --data_names=mnist
# View tail of output log files
tail data/results/data-mnist_model-deep-copula_n_jobs-1.log
# Command for all models and datasets (using commas to separate)
python scripts/icml_2018_experiment.py --model_names=deep-copula,image-pairs-copula,image-pairs-tree --data_names=mnist,cifar10
# Command to run all experiments in parallel using subprocesses
python scripts/icml_2018_experiment.py --model_names=deep-copula,image-pairs-copula,image-pairs-tree --data_names=mnist,cifar10 --parallel_subprocesses=True Please read through the following high-level guidelines:
scikit-learn coding guidelines -
http://scikit-learn.org/stable/developers/contributing.html#coding-guidelinesFor this particular project, please follow these additional guidelines:
Use lower case with underscores for variable names and functions.
Please use longer names with full spellings especially for public
interfaces to allow for super lightweight documentation. The variable
names should be descriptive of its function. For example, a
constructor name should be fitted_canonical_destructor rather
than fitted_destructor or destructor or fit_canon_destr
or fcd. Another example, univariate_estimators rather than
univ_est or univariate_est or uest. It is much easier to
change a long variable name to short one than the other way around.
Methods should generally be private designated by underscore prefix unless sure the method should be exposed publicly.
For non-negative integer count variables prefix with n_ rather
than num_ or number_of_
Use variable names n_samples, n_features, and
n_components (number of mixture components, number of PCA
vectors, etc) and n_layers instead of ambiguous single letter
variable names like n, p or k.
In the library and tests, please use the logging API instead of print
statements. In particular, create a logger for each module and call
the appropriate logging function (usually logger.debug(message))
.. code:: python
import logging logger = logging.getLogger(name) def foo(): logger.debug('Checking inside foo')
To avoid the module from outputing anything unless requested, the
root module file __init__.py redirects the logging output to
None as follows:
.. code:: python
import logging from logging import NullHandler logging.getLogger(name).addHandler(NullHandler())
Thus, to view these logs when executing a program and capture warnings as logs for a particular module you must setup logging to output to standard out (and/or a local file). For example, you could write:
.. code:: python
logging.basicConfig(stream=sys.stdout) # Push towards stdout instead of null handler logging.captureWarnings(True) # Capture warnings in loggers logging.getLogger('ddl').setLevel(logging.DEBUG) # Show everything above DEBUG level for the root ddl module