enijkamp

# Map Cooperative Networks

import shutil
import csv
from map_code.coop_elm import *
import argparse
from acd.models import _netE, _netG, _netI

parser = argparse.ArgumentParser()

enijkamp

t_pl = tf.placeholder(shape=[], dtype=tf.float32, name='t_pl')
# B(t) = (1-t)^3 P0 + 3(1-t)^2 t P1 + 3(1-t) t^2 P2 + t^3 P3
wt = [((1-t_pl) ** 3)*np.array(P0[i]) + (3*(1-t_pl)**2)*t_pl*P1[i] + (3*(t_pl-1)*(t_pl**2))*P2[i] + (t_pl**3)*P3[i] for i in range(len(P0))]
model_t = Cifar10ModelSimple(wt, resnet_size=32)
logits_t = model_t(x, is_training)
cross_entropy_t = tf.losses.softmax_cross_entropy(logits=logits_t, onehot_labels=y_)
correct_prediction_t = tf.equal(tf.argmax(logits_t, axis=1), tf.argmax(y_, axis=1))
accuracy_t = tf.reduce_mean(tf.cast(correct_prediction_t, tf.float32))

distance_t = distance_to(wt, P0)

enijkamp

\begin{table}[h]
	\begin{tabular}{|>{\raggedright}m{0.125\textwidth}|>{\raggedright}m{0.56\textwidth}|>{\centering}m{0.08\textwidth}|>{\centering}m{0.07\textwidth}|>{\centering}m{0.07\textwidth}|}
		\hline 
		Project & Experiment & CPU SU's & GPU SU's & Storage (GB)\tabularnewline
		\hline 
		\hline 
		\multirow{3}{0.15\textwidth}{(1) Learning Generative Models}
		& (a) Single-Grid Minimal Contrastive Divergence & $2,000$ & $10,000$ & $100$\tabularnewline
		\cline{2-5} 
		& (b) Latent Variable Convolutional Energy-based Models & $2,000$ & $12,500$ & $200$\tabularnewline

enijkamp

sudo singularity build --sandbox ubuntu/ docker://ubuntu:16.04

singularity shell ubuntu

apt update
apt install wget cron
wget https://d2t3ff60b2tol4.cloudfront.net/builds/insync-headless_1.5.2.37346-wheezy_amd64.deb
dpkg -i insync-headless_1.5.2.37346-wheezy_amd64.deb

singularity build ubuntu.img ubuntu

enijkamp

  case 'log' % Laplacian of Gaussian
     % first calculate Gaussian
     siz   = (p2-1)/2;
     std2   = p3^2;
     
     [x,y] = meshgrid(-siz(2):siz(2),-siz(1):siz(1));
     arg   = -(x.*x + y.*y)/(2*std2);
     
     h     = exp(arg);
     h(h<eps*max(h(:))) = 0;

enijkamp

   -0.1068   -0.1761   -0.1068
    0.2136    0.3522    0.2136
   -0.1068   -0.1761   -0.1068

enijkamp

#!/bin/bash
#SBATCH -N 1
#SBATCH -p GPU-shared
#SBATCH --ntasks-per-node 4
#SBATCH --gres=gpu:p100:1
#SBATCH -t 47:00:00

module load python2/2.7.11_gcc
module load cuda/8.0
cd

enijkamp

20900 en_gen_0=    -2.006107 en_gen_1=    -2.008458 en_gen_d=    -0.002351
20900 en_pos=[   -4.6743] en_neg=[   -1.9059]
20900 en_pos_test=[   -4.3532] en_neg_test=[   -2.4173]
20900 en_pos=[   -4.6743    -4.6741     0.0001] en_neg=[   -1.9059    -1.9036     0.0023]
20900 loss= -2.0061
20900 neg_range=[    0.0000     0.9999] neg_abs_sum=  239.6976
x_pos_0 = 3 - 3 + 2 + 2      
x_pos_0 = 3 - 3 + 2 + 2      
x_neg_0 = ( 1 * ( 3 3 3      
x_neg_1 = ( ( z ( + 3 3 )    

enijkamp

import torch.utils
import torch.nn.utils
from torch.utils import data

from torchvision import transforms

from PIL import Image, ImageDraw
import numpy as np

enijkamp

import torch.utils
import torch.nn.utils
from torch.utils import data

from torchvision import transforms

from PIL import Image, ImageDraw
import numpy as np