Add files via upload

Author: fiorezhang

picross_dataset.py

@@ -5,11 +5,11 @@
 from picross_generator import generator as gen
 
 def load_data(row, col, visible, num):
-	x = np.zeros((num, row, col), dtype='int16')
-	y = np.zeros((num, 2, row, col), dtype='int16')
+	x = np.zeros((num, 2, row, col), dtype='int16')
+	y = np.zeros((num, row, col), dtype='int16')
 
 	for i in range(num):
-		x[i], y[i, 0], y[i, 1] = gen(row, col, visible)
+		y[i], x[i, 0], x[i, 1] = gen(row, col, visible)
 
 	return x, y
 

picross_rnn.py

@@ -0,0 +1,98 @@
+#picross_rnn.py
+#copyright: fiorezhang@sina.com
+
+import numpy as np
+from keras.models import Sequential
+from keras import layers
+from picross_dataset import load_data as load
+
+#global variables for dataset
+SIZE_ROW = 3
+SIZE_COL = 5
+VISIBLE = 0.6
+TRAIN_SIZE = 5000
+TEST_SIZE = 500
+
+#generate data from picross generator/dataset functions
+print('-'*50)
+print('Generating data...')
+print('Picross size: ', SIZE_ROW, 'x', SIZE_COL, ', visible: ', VISIBLE*100, '%')
+print('Train samples: ', TRAIN_SIZE, ', test samples: ', TEST_SIZE)
+x, y = load(SIZE_ROW, SIZE_COL, VISIBLE, TRAIN_SIZE+TEST_SIZE)
+x_train, x_test = x[:TRAIN_SIZE], x[TRAIN_SIZE:]
+y_train, y_test = y[:TRAIN_SIZE], y[TRAIN_SIZE:]
+print(x.shape)
+print(y.shape)
+print(x_train.shape)
+print(y_train.shape)
+print(x[0])
+print(y[0])
+
+#set parameters for RNN modle
+RNN=layers.SimpleRNN
+HIDDEN_SIZE = 128
+BATCH_SIZE = 128
+LAYERS = 1
+ITERATION = 100
+EPOCHS = 10
+
+#build the model
+print('-'*50)
+print('Building model...')
+model = Sequential()
+#reshape input, connect 2 matrix and flatten
+model.add(layers.Reshape((2*SIZE_ROW*SIZE_COL, 1), input_shape=(2, SIZE_ROW, SIZE_COL)))
+#print(model.output_shape)
+#import data to a RNN
+model.add(RNN(HIDDEN_SIZE, dropout=0.1))
+#print(model.output_shape)
+#repeat r*c times for output
+model.add(layers.RepeatVector(SIZE_ROW*SIZE_COL))
+#print(model.output_shape)
+#rnn again, set return_sequences to output for all time pieces
+for _ in range(LAYERS):
+	model.add(RNN(HIDDEN_SIZE, return_sequences=True, dropout=0.1))
+#print(model.output_shape)
+#flatten, 3D->2D
+model.add(layers.TimeDistributed(layers.Dense(1)))
+#print(model.output_shape)
+#finally binary cross entropy
+model.add(layers.Activation('sigmoid'))
+#print(model.output_shape)
+#model.add(layers.Reshape((SIZE_ROW, SIZE_COL)))
+model.compile(loss='binary_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
+#print(model.output_shape)
+model.summary()
+
+#train the model and print information during the process
+print('-'*50) 
+print('Training...')
+for iteration in range(1, ITERATION):
+	print()
+	print('-'*50)
+	print('Iteration', iteration)
+	model.fit(x_train, y_train.reshape(-1, SIZE_ROW*SIZE_COL, 1),
+          batch_size=BATCH_SIZE,
+          epochs=EPOCHS,
+          validation_data=(x_test, y_test.reshape(-1, SIZE_ROW*SIZE_COL, 1)))	
+	#show result in the middle
+	for i in range(1): 
+		ind = np.random.randint(0, len(x_test))
+		rowx, rowy = x_test[np.array([ind])], y_test[np.array([ind])]
+		preds = model.predict_classes(rowx, verbose=0)
+		question = rowx[0]
+		#print(question.shape)
+		correct = rowy[0]
+		#print(correct.shape)
+		#print(preds.shape)
+		guess = preds[0].reshape(SIZE_ROW, SIZE_COL)	
+		print('Q','- '*25)
+		print(question)
+		print('A','- '*25)
+		print(correct)
+		if (correct == guess).all():
+			print('Y','- '*25)
+		else:
+			print('N','- '*25)
+		print(guess)
+		print('- '*25)