Matt Shaffer planetceres

UC Berkeley http://www.discovermatt.com

planetceres / tf_1-8_keras_2-1_gpu_setup.md

Created July 23, 2018 06:11

Provision Ubuntu 16.04 with CUDA 9.0 CUDNN 7.1.4 Tensorflow 1.8.0 Keras 2.1.5

Requirements

planetceres / horovod_cluster_setup.md

Last active March 5, 2020 20:56

Horovod Cluster Setup with CUDA 9.0, Miniconda3 and Tensorflow 1.8rc

Update CentOS and reboot (you will need to login again):

sudo yum update -y
sudo reboot

planetceres / conda_remove_all.sh

Last active March 27, 2018 21:28

Remove all conda environmets 'bash conda_remove_all.sh'

	#!/bin/bash

	. ~/anaconda3/etc/profile.d/conda.sh

	ENVS=$(conda env list \| grep '^\w' \| cut -d' ' -f1)
	for env in $ENVS; do
	conda remove -n $env --all
	echo "Removing $env"
	done

planetceres / conda_export_all.sh

Last active March 27, 2018 21:27

Export all conda env to environment.yml `bash conda_export_all.sh <folder path for backup>`

	#!/bin/bash

	. ~/anaconda3/etc/profile.d/conda.sh

	NOW=$(date "+%Y-%m-%d")
	if [ $1 -eq 0 ]; then
	DIR='~/Desktop/condas/'
	else
	DIR=$1
	fi

planetceres / conda_remove_all.sh

Last active March 27, 2018 21:27

[Deprecated] Remove all Anaconda environments with prompt `bash conda_remove_all.sh`

	#!/bin/bash

	ENVS=$(conda env list \| grep '^\w' \| cut -d' ' -f1)
	for env in $ENVS; do
	conda remove -n $env --all
	echo "Removing $env"
	done

planetceres / conda_backup_env.sh

Last active March 27, 2018 21:27

[Deprecated] Backup all Anaconda environments `bash conda_backup_env.sh <folder path for backup>`

	#!/bin/bash

	NOW=$(date "+%Y-%m-%d")
	DIR=$1
	mkdir $DIR/envs-$NOW
	ENVS=$(conda env list \| grep '^\w' \| cut -d' ' -f1)
	for env in $ENVS; do
	source activate $env
	conda env export > $DIR/envs-$NOW/$env.yml
	echo "Exporting $env to .yml"

planetceres / adaboost.py

Created September 24, 2016 00:00 — forked from tristanwietsma/adaboost.py

AdaBoost Python implementation of the AdaBoost (Adaptive Boosting) classification algorithm.

	from __future__ import division
	from numpy import *

	class AdaBoost:

	def __init__(self, training_set):
	self.training_set = training_set
	self.N = len(self.training_set)
	self.weights = ones(self.N)/self.N
	self.RULES = []