stefanthaler’s gists

stefanthaler / group_trunc_seq_length.py

Created July 22, 2020 11:26

Tensorflow DataSet transformation that groups sequential data into buckets and truncates them instead of padding.

	#
	import tensorflow as tf
	import numpy as np

	from tensorflow.python.framework import ops
	from tensorflow.python.data.ops import dataset_ops
	from tensorflow.python.util.tf_export import tf_export
	from tensorflow.python.data.experimental import group_by_window

	from tensorflow.python.framework import constant_op

stefanthaler / python_xlib_select_spike.py

Created August 31, 2019 19:40

A small example for python-xlib that works of events in a semi-blocking way using select.

	#!/usr/bin/python3

	import sys
	import os
	from Xlib import X, display, Xutil
	import select

	# http://python-xlib.sourceforge.net/doc/html/python-xlib_11.html

	# Application window (only one)

stefanthaler / ensure_gradient_flow.py

Created October 11, 2017 11:53

Simple method for ensuring that the gradient flows through all your variables in TensorFlow.

	import tensorflow as tf # 1.3.0
	import numpy as np

	# model + input
	w = tf.Variable(1, name="w", dtype=tf.float32 ) # parameter to optimize for
	x = tf.placeholder(shape=(), dtype=tf.float32, name="x") # input

	# graph operations
	diff_op = tf.multiply(w, x) # an operation that is differentiable
	non_diff_op = tf.cast(tf.equal(diff_op, 0), dtype=tf.float32) # operation that is non differentiable

stefanthaler / check_reference_links.py

Created April 24, 2017 12:53

Checks a latex .bib file for links. If the script finds links, it tries to reach them to see if they are still alive.

	# needs bibtexparser==0.6.2, argparse
	"""
	Parses a Bib file and checks if all links are available
	"""

	# imports
	import bibtexparser
	import argparse
	import re
	import requests

stefanthaler / tensorflow-dynamic-rnn-explained.ipynb

Created February 22, 2017 16:13

Tensorflow Dynamic RNN explained

Sorry, something went wrong. Reload?

Sorry, we cannot display this file.

Sorry, this file is invalid so it cannot be displayed.

stefanthaler / colorize_word_embeddings.py

Created February 3, 2017 14:38

A simple example to demonstrate how to link embedding metadata to word embeddings in tensorflow / tensorboard

	"""
	Simple example to demostrate the embedding visualization for word embeddings in tensorflow / tensorboard
	https://www.tensorflow.org/how_tos/embedding_viz/
	"""

	import tensorflow as tf
	import os
	assert tf.__version__ == '1.0.0-rc0' # if code breaks, check tensorflow version
	from tensorflow.contrib.tensorboard.plugins import projector

stefanthaler / lstm_backwards_embedding.py

Created December 1, 2016 15:56

Keras Embedding + Backwards LSTM

	from keras.models import Sequential # model used
	from keras.layers import Dense, Embedding, LSTM # layers used

	batch_size = 2 # how many sequence to process in parallel
	time_steps = 3 # lstm length, number of cells, etc.
	input_dim = 1 # number of features
	embedding_size = 5 # size of embedding vector

	model = Sequential()
	model.add(Embedding(

stefanthaler / stateful_lstm_embedding.py

Last active September 2, 2020 14:10

Simple example for a stateful keras LSTM with embedding.

	"""
	Learning Task:
	Given a sequence, predict a label based on the first value of the sequence

	Explanation of stateful LSTM and setup:
	http://philipperemy.github.io/keras-stateful-lstm/

	Exmple:
	given a sequence [1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], predict 1
	given a sequence [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], predict 0

stefanthaler / fischer_wagner_explained.py

Created November 23, 2016 13:12

Implementation of the Fisher Wagner algorithm for calculation of the Levensthein distance which outputs all steps to enable understanding.

	import numpy as np
	from scipy.stats import entropy # calculate kl divergence #http://scipy.github.io/devdocs/generated/scipy.stats.entropy.html
	np.set_printoptions(precision=2)
	# https://en.wikipedia.org/wiki/Wagner%E2%80%93Fischer_algorithm
	# Original Paper http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.367.5281&rep=rep1&type=pdf

	deletion_cost = 1
	insertion_cost = 1
	substitution_cost = 1

stefanthaler / nutritients_lp.py

Last active July 3, 2016 14:34

	# sudo pip install pulp
	# sudo apt-get install python-glpk gplk-utils
	from pulp import *

	optimization_problem = LpProblem("Nutrients Optimization", LpMinimize)

	# Cost Parameters for Selecting a food
	# diversity (minimum amount of certain foods)
	# diversity (minimum amount of vegetables)
	# multiple cost factors (money, co2, water)

Stefan Thaler stefanthaler