forked from arrdem/OpenSourcerer
-
Notifications
You must be signed in to change notification settings - Fork 0
/
land_net_trainer.py
62 lines (50 loc) · 1.82 KB
/
land_net_trainer.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
#!/usr/bin/env python2
# -*- coding: utf-8 -*-
from ConfigParser import SafeConfigParser
import McKenzieLib.learning.neuralnet as netlib
import lib.ai.land_ai as ai
import pickle
import sys
__lands__ = ['Island', 'Swamp', 'Mountain', 'Plains', 'Forest']
__cost_types__ = ['Blue', 'Black', 'Red', 'White', 'Green']
def main():
# Land Filler Network Spec
# Inputs (conceptual):
# Most expensive card, Average cc, percentage split by color,
# deck size without lands
# Order
# 'Blue', 'Black', 'Red', 'White', 'Green'
# Inputs (practical):
# 5*float for the most expensive card //OLD
# 5*float for the average cc
# 5*float for percentage split by color
# 1*float for card count sans lands
# Total: 16 inputs
#
# Outputs:
# 5*float, one per card type (count of reccomended)
parser = SafeConfigParser()
parser.read('settings.ini')
ds = []
for l in open(parser.get('land_net', 'corpus')):
try:
i, o = map(ai.import_vector, l.split('|'))
if i and o:
print i, o
ds.append((i, o,))
except:
pass
net = None
if '-l' in sys.argv:
net = pickle.load(open(parser.get('land_net', 'brain')))
else:
net = netlib.NN(int(parser.get('land_net', 'insize')),
int(parser.get('land_net', 'hidden')),
int(parser.get('land_net', 'outsize')))
print "CREATED NET WITH DIMENSIONS %i %i %i" % (net.ni, net.nh, net.no)
print "STARTING TRAINING....."
net.train(ds, iterations=int(parser.get('land_net', 'iterations')))
f = open(parser.get('land_net', 'brain'),'wb')
pickle.dump(net, f, 2)
if __name__ == '__main__':
main()