In [2]:
import theano
import sys, os
sys.path.insert(1, os.path.join('../utils'))
from utils import *
path = get_file('nietzsche.txt', origin="https://s3.amazonaws.com/text-datasets/nietzsche.txt")
text = open(path).read()
print('corpus length:', len(text))
chars = sorted(list(set(text)))
vocab_size = len(chars) + 1
print('total chars:', vocab_size)
chars.insert(0, "\0")
char_indices = dict((c, i) for i, c in enumerate(chars))
indices_char = dict((i, c) for i, c in enumerate(chars))
idx = [char_indices[c] for c in text]
n_fac = 42
n_hidden = 256
cs = 8
c_in_dat = [[idx[i+n] for i in xrange(0, len(idx)-1-cs, cs)] for n in range(cs)]
c_out_dat = [[idx[i+n] for i in xrange(1, len(idx)-cs, cs)] for n in range(cs)]
xs = [np.stack(c[:-2]) for c in c_in_dat]
ys = [np.stack(c[:-2]) for c in c_out_dat]
# before model fit
[xs[n][:cs] for n in range(cs)]
Out[2]:
In [3]:
model = Sequential([
SimpleRNN(n_hidden, return_sequences=True, input_shape=(cs, vocab_size),
activation='relu', inner_init='identity'),
TimeDistributed(Dense(vocab_size, activation='softmax')),
])
model.compile(loss='categorical_crossentropy', optimizer=Adam())
# no embedding layer, so inputs must be onhotted
oh_ys = [to_categorical(o, vocab_size) for o in ys]
oh_y_rnn = np.stack(oh_ys, axis=1)
oh_xs = [to_categorical(o, vocab_size) for o in xs]
oh_x_rnn = np.stack(oh_xs, axis=1)
oh_x_rnn.shape, oh_y_rnn.shape
Out[3]:
In [5]:
model.fit(oh_x_rnn, oh_y_rnn, batch_size=64, nb_epoch=1)
Out[5]:
In [6]:
# after model fit
[xs[n][:cs] for n in xrange(cs)]
Out[6]:
In [7]:
# THEANO RNN
n_input = vocab_size
n_output = vocab_size
def init_wgts(rows, cols):
scale = math.sqrt(2/rows)
return shared(normal(scale=scale, size=(rows,cols)).astype(np.float32))
def init_bias(rows):
return shared(np.zeros(rows, dtype=np.float32))
def wgts_and_bias(n_in, n_out):
return init_wgts(n_in, n_out), init_bias(n_out)
def id_and_bias(n):
return shared(np.eye(n, dtype=np.float32)), init_bias(n)
# Theano Variables
t_inp = T.matrix('inp')
t_outp = T.matrix('outp')
t_h0 = T.vector('h0')
lr = T.scalar('lr')
all_args = [t_h0, t_inp, t_outp, lr]
W_h = id_and_bias(n_hidden)
W_x = wgts_and_bias(n_input, n_hidden)
W_y = wgts_and_bias(n_hidden, n_output)
w_all = list(chain.from_iterable([W_h, W_x, W_y]))
def step(x, h, W_h, b_h, W_x, b_x, W_y, b_y):
# Calculate the hidden activations
h = nnet.relu(T.dot(x, W_x) + b_x + T.dot(h, W_h) + b_h)
# Calculate the output activations
y = nnet.softmax(T.dot(h, W_y) + b_y)
# Return both (the `Flatten()` is to work around a Theano bug)
return h, T.flatten(y, 1)
[v_h, v_y], _ = theano.scan(step, sequences=t_inp,
outputs_info=[t_h0, None], non_sequences=w_all)
error = nnet.categorical_crossentropy(v_y, t_outp).sum()
g_all = T.grad(error, w_all)
def upd_dict(wgts, grads, lr):
return OrderedDict({w: w - g * lr for (w,g) in zip(wgts, grads)})
upd = upd_dict(w_all, g_all, lr)
fn = theano.function(all_args, error, updates=upd, allow_input_downcast=True)
X = oh_x_rnn
Y = oh_y_rnn
print X.shape, Y.shape
err = 0.0; l_rate = 0.01
for i in xrange(len(X)):
err += fn(np.zeros(n_hidden), X[i], Y[i], l_rate)
if i % 1000 == 999:
print ("ErrorX:{:.3f}".format(err/1000))
err=0.0
Looks like this model fit isn't the one that changes x....
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]: