def gbootstrap(ginput, goutput, fnames):
print "Bootstrapping", ginput.getName(), "with",
if fnames == []:
print "basic elements"
else:
for fname in fnames:
print "\""+fname+"\"",
print ""
# fixed parameters
spectral_radius = 0.9
# Sequences
seqs = []
for fname in fnames:
x = aux.load_words_from_text(fname, ginput)
seqs.extend(x)
if fnames == []:
seqs = list(goutput.alphabet) # to avoid using a reference!
seqs.remove(goutput.stop_symbol)
#print seqs
#print "words #", len(words)
train_set = seqs
#test_set = ???
to_train = []
#to_test = []
for seq in train_set:
if "SeqWord" in ginput.getName():
syls = ginput.getSyllables(seq)
for j in range(len(syls)-1):
to_train.append(list("".join(syls[0:j])))
to_train.extend(list(syls))
to_train.append(list(seq) + [ginput.stop_symbol])
else:
to_train.append([seq , ginput.stop_symbol])
wvectorizer = WVectorizer.WVectorizer(ginput.getSize(), ginput.getInternalSize(), spectral_radius)
#to_train = list(set(to_train))
#print to_train
if "" in to_train:
to_train.remove("")
if [] in to_train:
to_train.remove([])
#print to_train
to_train = map(lambda x: (x, wvectorizer.wvectorize(ginput.input(x))), to_train)
#to_test = map(lambda x: (x, wvectorizer.wvectorize(winput.input(x))), to_test)
#print to_train
wunvectorizer = WunVectorizer.WunVectorizer(spectral_radius, goutput)
wunvectorizer.train(to_train, [], True)
#print ""
return wvectorizer, wunvectorizer
def gbootstrap(ginput, goutput, fnames):
print "Bootstrapping", ginput.getName(), "with",
if fnames == []:
print "basic elements"
else:
for fname in fnames:
print "\"" + fname + "\"",
print ""
# fixed parameters
spectral_radius = 0.9
# Sequences
seqs = []
for fname in fnames:
x = aux.load_words_from_text(fname, ginput)
seqs.extend(x)
if fnames == []:
seqs = list(goutput.alphabet) # to avoid using a reference!
seqs.remove(goutput.stop_symbol)
#print seqs
#print "words #", len(words)
train_set = seqs
#test_set = ???
to_train = []
#to_test = []
for seq in train_set:
if "SeqWord" in ginput.getName():
syls = ginput.getSyllables(seq)
for j in range(len(syls) - 1):
to_train.append(list("".join(syls[0:j])))
to_train.extend(list(syls))
to_train.append(list(seq) + [ginput.stop_symbol])
else:
to_train.append([seq, ginput.stop_symbol])
wvectorizer = WVectorizer.WVectorizer(ginput.getSize(),
ginput.getInternalSize(),
spectral_radius)
#to_train = list(set(to_train))
#print to_train
if "" in to_train:
to_train.remove("")
if [] in to_train:
to_train.remove([])
#print to_train
to_train = map(lambda x: (x, wvectorizer.wvectorize(ginput.input(x))),
to_train)
#to_test = map(lambda x: (x, wvectorizer.wvectorize(winput.input(x))), to_test)
#print to_train
wunvectorizer = WunVectorizer.WunVectorizer(spectral_radius, goutput)
wunvectorizer.train(to_train, [], True)
#print ""
return wvectorizer, wunvectorizer
Copyright 2010, 2011, 2012 by neuromancer
"""
import src.io.text.English as English
import src.core.Mind as Mind
import src.Stats as Stats
from src.aux import load_words_from_text
# mind initilization
winput = English.InputWords()
woutput = English.OutputWords()
mind = Mind.Mind([winput], [woutput], ["data/pos/categories.txt"])
cats_files = load_words_from_text("data/pos/categories.longer.txt", winput)
cats = load_words_from_text("data/pos/categories.txt", winput)
stats = Stats.Stats(mind)
# mind training
for (cat, cat_file) in zip(cats, cats_files):
print "Assimilating", cat_file + "s.."
words = load_words_from_text("data/pos/" + cat_file + "s_train.txt",
winput)
cat_inputs = dict(SeqWordsEn=list(cat))
for word in words:
Copyright 2010, 2011, 2012 by neuromancer
"""
import src.io.text.English as English
import src.core.Mind as Mind
import src.Stats as Stats
from src.aux import load_words_from_text
# mind initilization
winput = English.InputWords()
woutput = English.OutputWords()
mind = Mind.Mind([winput], [woutput], ["data/pos/categories.txt"])
cats_files = load_words_from_text("data/pos/categories.longer.txt", winput)
cats = load_words_from_text("data/pos/categories.txt", winput)
stats = Stats.Stats(mind)
# mind training
for (cat,cat_file) in zip(cats,cats_files):
print "Assimilating", cat_file+"s.."
words = load_words_from_text("data/pos/"+cat_file+"s_train.txt", winput)
cat_inputs = dict( SeqWordsEn = list(cat))
for word in words:
word_inputs = dict( SeqWordsEn = list(word))