lib/exercises-in-programming-style
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Clean up

Browse Source
This commit is contained in:
Crista Lopes
2019-12-31 10:17:12 -08:00
parent 9c51b2fc1c
commit a8379345c7
1 changed files with 2 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
38-rnn/tf-38.py
View File

@@ -1,9 +1,5 @@
from keras.models import Sequential from keras.models import Sequential
from keras.layers import Dense, LSTM, RepeatVector, TimeDistributed from keras.layers import Dense, LSTM
from keras.losses import binary_crossentropy, categorical_crossentropy
from keras.optimizers import SGD
from keras. metrics import top_k_categorical_accuracy
from keras import backend as K
import numpy as np import numpy as np
import sys, os, string, random import sys, os, string, random
@@ -14,13 +10,9 @@ indices_char = dict((i, c) for i, c in enumerate(characters))
INPUT_VOCAB_SIZE = len(characters) INPUT_VOCAB_SIZE = len(characters)
BATCH_SIZE = 200 BATCH_SIZE = 200
HIDDEN_SIZE = 100 HIDDEN_SIZE = 100
N_LAYERS = 1
TIME_STEPS = 3 TIME_STEPS = 3
def encode_one_hot(line): def encode_one_hot(line):
# remain = len(line) % TIME_STEPS
# if remain != 0:
# line = line + ' ' * (TIME_STEPS-remain)
line = " " + line line = " " + line
x = np.zeros((len(line), INPUT_VOCAB_SIZE)) x = np.zeros((len(line), INPUT_VOCAB_SIZE))
for i, c in enumerate(line): for i, c in enumerate(line):
@@ -40,8 +32,6 @@ def prepare_for_rnn(x):
ind = [np.array(np.arange(i, i+TIME_STEPS)) for i in range(x.shape[0] - TIME_STEPS + 1)] ind = [np.array(np.arange(i, i+TIME_STEPS)) for i in range(x.shape[0] - TIME_STEPS + 1)]
ind = np.array(ind, dtype=np.int32) ind = np.array(ind, dtype=np.int32)
x_rnn = x[ind] x_rnn = x[ind]
# np.append(x_rnn, np.zeros((TIME_STEPS, INPUT_VOCAB_SIZE)), axis=0)
# np.append(x_rnn, np.zeros((TIME_STEPS, INPUT_VOCAB_SIZE)), axis=0)
return x_rnn return x_rnn
def input_generator(nsamples): def input_generator(nsamples):
@@ -63,8 +53,6 @@ def input_generator(nsamples):
while True: while True:
input_data, expected = generate_line() input_data, expected = generate_line()
#print("Input :", input_data)
#print("Output:", expected)
data_in = encode_one_hot(input_data) data_in = encode_one_hot(input_data)
data_out = encode_one_hot(expected) data_out = encode_one_hot(expected)
yield prepare_for_rnn(data_in), data_out yield prepare_for_rnn(data_in), data_out
@@ -83,12 +71,7 @@ def train(model):
def build_model(): def build_model():
model = Sequential() model = Sequential()
r_layer = LSTM(HIDDEN_SIZE, input_shape=(None, INPUT_VOCAB_SIZE)) model.add(LSTM(HIDDEN_SIZE, input_shape=(None, INPUT_VOCAB_SIZE)))
model.add(r_layer)
# model.add(RepeatVector(TIME_STEPS))
# for _ in range(N_LAYERS):
# model.add(LSTM(HIDDEN_SIZE, return_sequences=True))
# model.add(TimeDistributed(Dense(INPUT_VOCAB_SIZE, activation='softmax')))
model.add(Dense(INPUT_VOCAB_SIZE, activation='softmax')) model.add(Dense(INPUT_VOCAB_SIZE, activation='softmax'))
return model return model