f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Demonstrates that cuDNN does not choose the fastest code path internally.

It first asks cuDNN to compute a particular full convolution as a forward
pass with sufficient zero padding of the input.
It then asks cuDNN to compute the same as a backward pass of a valid convolution
without zero padding of the input.

f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Strips the pHYs chunk (resolution information) of a .png file.
Handy if pdflatex fails with "dimension too large" due to the
PNG resolution being set to 1.

For usage information, call without any parameters.

f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Example employing Lasagne for digit generation using the MNIST dataset and
Wasserstein Generative Adversarial Networks
(WGANs, see https://arxiv.org/abs/1701.07875 for the paper and
https://github.com/martinarjovsky/WassersteinGAN for the "official" code).

It is based on a DCGAN example:

f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Example employing Lasagne for digit generation using the MNIST dataset and
Least Squares Generative Adversarial Networks
(LSGANs, see https://arxiv.org/abs/1611.04076 for the paper).

It is based on a WGAN example:
https://gist.github.com/f0k/f3190ebba6c53887d598d03119ca2066

f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Example employing Lasagne for digit generation using the MNIST dataset and
Deep Convolutional Generative Adversarial Networks
(DCGANs, see http://arxiv.org/abs/1511.06434).

It is based on the MNIST example in Lasagne:
http://lasagne.readthedocs.org/en/latest/user/tutorial.html

f0k

# -*- coding: utf-8 -*-

"""
PCEN Layer for Lasagne.

Author: Jan Schlüter, OFAI
"""

import numpy as np
import theano

f0k

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

"""
Merges and deduplicates one or more .ics / ical files.

For usage information, call with --help.

Author: Jan Schlüter
"""

f0k

# -*- coding: utf-8 -*-

"""
Preliminary implementation of batch normalization for Lasagne.
Does not include a way to properly compute the normalization factors over the
full training set for testing, but can be used as a drop-in for training and
validation.

Author: Jan Schlüter
"""

f0k

#!/bin/bash
# Sets each CUDA device to persistence mode and sets the application clock
# and power limit to the device's maximum supported values.
# When run with "--dry-run" as first command line argument or not as superuser,
# will display the commands, otherwise it will execute them.
#
# Hint: To run this at boot time, place this script in /root and create a file
# /etc/cron.d/nvidia_boost with the following single line:
# @reboot root /root/nvidia_boost.sh >/dev/null
#

f0k

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""
Compares some algorithms for computing a sliding median of an array.

Results:
  scipy.signal.medfilt2d is a bit faster than scipy.ndimage.filter.median_filter
  and significantly faster than scipy.signal.medfilt.
  Maintaining a sorted list of the window becomes faster than that for a filter