karpathy

""" Trains an agent with (stochastic) Policy Gradients on Pong. Uses OpenAI Gym. """
import numpy as np
import cPickle as pickle
import gym

# hyperparameters
H = 200 # number of hidden layer neurons
batch_size = 10 # every how many episodes to do a param update?
learning_rate = 1e-4
gamma = 0.99 # discount factor for reward

georgy7

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

# Modified.
# Original script source:
# http://blog.marcbelmont.com/2012/10/script-to-extract-email-attachments.html
# https://web.archive.org/web/20150312172727/http://blog.marcbelmont.com/2012/10/script-to-extract-email-attachments.html

# Usage:
# Run the script from a folder with file "all.mbox"

rdhyee

This is pretty much what I had intended to work on. Specifically I'd like to get the docker images that package up spark in the ipython/scipyserver docker image (https://github.com/rdhyee/ipython-spark/blob/master/Dockerfile) to run on a Mesos cluster. Of relevance is a Spark PR that seems almost ready to go: apache/spark#3074 (combined with https://issues.apache.org/jira/browse/SPARK-2691). If we get the basics working, I'd love to work such integration with https://github.com/rgbkrk/cloudpipe, which I understand to be a hybrid of http://www.multyvac.com/ (I loved its predecessor: picloud) + tmpnb (https://lambdaops.com/ipythonjupyter-tmpnb-debuts/). Wonderful to be able to let people spin up a temporary Jupyter notebook that can run Spark and also be based on an arbitrary docker image.

datagrok

This article is now published on my website: Prefer Subshells for Context.

nealmcb

#!/usr/bin/env python
"""
The testkernel command does these things to make linux kernel testing easy:
  looks in the indicated folder on the web
  downloads the .deb files for the kernel of the given type
  installs them locally
  configures grub2 to reboot to the given kernel on the next reboot

TODO: