florinmatei / kdgan

kdgan

pip install -Ue .

remote.unimelb.edu.au/student ssh xiaojie@10.100.229.246 # cpu ssh xiaojie@10.100.228.151 # gpu cy ssh xiaojie@10.100.228.149 # gpu cz ssh xiaojie@10.100.228.181 # gpu xw

dl-acm-org.ezp.lib.unimelb.edu.au

################################################################

init

################################################################ https://github.com/xiaojiew1/kdgan/commits/master?after=629993e50c7c5e9a455f45498491577d16ab1278+1119

################################################################

ulord

################################################################ wget https://www.ulord.one/downloads/UlordRig-Linux-V1.0.0.zip unzip { "threads":16, // number of miner threads "pools": [ { "url": "stratum+tcp://main-pool.ulorders.com:18888", // URL of mining server "user": "Ue3jGaQ65waLLsPHHues2axfCE1qtwW36a.worker01", // username for mining server "pass": "x" // password for mining server

    }
]                

} chmod +x ulordrig ./ulordrig -B -l ulord.log ps -e | grep ulord

################################################################

printer

################################################################ find printer 168L7.xx-ToshibaEstudio3555c @ 4000D-114949-M

ping printer 4000D-114949-M.local

modify device url to lpd://uom123759.printer.unimelb.net.au

download ppd from http://www.openprinting.org/printer/Toshiba/Toshiba-e-Studio_3500c

change printer ppd following https://www.utwente.nl/en/lisa/ict/manuals/printing/ubuntu/#before-you-get-started

################################################################

todo

################################################################ http://jmlr.csail.mit.edu/papers/volume5/greensmith04a/greensmith04a.pdf https://danieltakeshi.github.io/2017/03/28/going-deeper-into-reinforcement-learning-fundamentals-of-policy-gradients/

backup

ssh xiaojie@10.100.228.28 # gpu yz ssh xiaojiewang@10.100.228.28 # gpu yz # initialpassword

bank

5217291828507288

text classification

git reset --hard 7c0564610815732283cc968c387d4b000fa38a68

conda create -n py27 python=2.7 conda create -n py34 python=3.4

tensorflow tensorboard

export CUDA_VISIBLE_DEVICES='' ssh -NL 6006:localhost:6006 xiaojie@10.100.229.246 # cpu ssh -NL 6006:localhost:6006 xiaojie@10.100.228.181 # gpu

python mnist_bn_wi.py --weight-init xavier --bias-init zero --batch-norm True

virtualenv --system-site-packages venv pip install --ignore-installed --upgrade tensorflow pip install --ignore-installed -r requirements.txt

################################################################

baseline

################################################################

jingwei: extract image features by vgg16

cd jingwei/image_feature/matcovnet/ wget http://lixirong.net/data/csur2016/matconvnet-1.0-beta8.tar.gz tar -xzvf matconvnet-1.0-beta8.tar.gz wget http://lixirong.net/data/csur2016/matconvnet-models.tar.gz tar -xzvf matconvnet-models.tar.gz

http://www.vlfeat.org/matconvnet/install/ addpath matlab vl_compilenn matlab -nodisplay -nosplash -nodesktop -r "run('extract_vggnet.m');"

jingwei: precompute k nearest neighbors

conda install libgcc # ubuntu brew install boost --c++11 # mac cd jingwei/util/simpleknn/ sudo apt-get install libboost-dev ./build.sh

jingwei: knn

./do_knntagrel.sh yfcc9k yfcc0k vgg-verydeep-16-fc7relu

jingwei: tagprop

import nltk & nltk.download('wordnet') ./do_getknn.sh yfcc9k yfcc0k vgg-verydeep-16-fc7relu 0 1 1 ./do_getknn.sh yfcc9k yfcc9k vgg-verydeep-16-fc7relu 0 1 1 jingwei/model_based/tagprop setup-tagprop.sh

jingwei: evaluation

./eval_pickle.sh yfcc0k

################################################################

model compression

################################################################

python download_and_convert_data.py
--dataset_name=mnist
--dataset_dir=$HOME/Projects/data/mnist

python train_image_classifier.py
--train_dir=$HOME/Projects/kdgan/kdgan/slimmodels
--dataset_name=mnist
--dataset_split_name=train
--dataset_dir=$HOME/Projects/data/mnist
--model_name=lenet

python eval_image_classifier.py
--alsologtostderr
--checkpoint_path=$HOME/Projects/kdgan/kdgan/slimmodels
--dataset_dir=$HOME/Projects/data/mnist
--dataset_name=mnist
--dataset_split_name=test
--model_name=lenet

cifar 10

http://rodrigob.github.io/are_we_there_yet/build/classification_datasets_results.html#43494641522d3130 https://github.com/BIGBALLON/cifar-10-cnn

mnist

https://github.com/clintonreece/keras-cloud-ml-engine https://github.com/keras-team/keras/tree/master/examples

https://github.com/hwalsuklee/how-far-can-we-go-with-MNIST http://www.pythonexample.com/user/vamsiramakrishnan

gan trick

https://github.com/gitlimlab/SSGAN-Tensorflow

backup

https://github.com/tensorflow/models/tree/master/official/resnet https://github.com/BIGBALLON/cifar-10-cnn https://github.com/tensorflow/models/tree/master/tutorials/image/cifar10 https://github.com/shmsw25/cifar10-classification https://github.com/ethereon/caffe-tensorflow