def main():
(waves, labels), (x_validate, y_validate) = load_data()
x_mean = 4.3854903e-05 # np.concatenate(waves).mean()
x_std = 0.042366702 # np.concatenate(waves).std()
waves = tuple(map(lambda wave: (wave - x_mean) / x_std, waves))
x_validate = (x_validate - x_mean) / x_std
model = Model(
*juxt(identity, computational_graph(max(y_validate) + 1))(Input(
shape=x_validate.shape[1:])))
model.compile(loss='sparse_categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
model.summary()
batch_size = 50
epoch_size = 500
results = model.fit_generator(
data_generator(waves, labels, batch_size),
steps_per_epoch=8000 // batch_size,
epochs=epoch_size,
validation_data=(x_validate, y_validate),
callbacks=[ReduceLROnPlateau(factor=0.5, patience=50, verbose=1)])
with open('./results/history.pickle', 'wb') as f:
pickle.dump(results.history, f)
save_model(model, './results/model.h5')
del model
def main():
(x_train, y_train), (x_validation, y_validation) = load_data()
model = Model(*juxt(identity, computational_graph(y_train.shape[1]))(Input(shape=x_train.shape[1:])))
model.compile(loss='categorical_crossentropy', optimizer=SGD(momentum=0.9), metrics=['accuracy'])
model.summary()
# plot_model(model, to_file='./results/model.png')
train_data = ImageDataGenerator(featurewise_center=True, featurewise_std_normalization=True, width_shift_range=0.125, height_shift_range=0.125, horizontal_flip=True)
validation_data = ImageDataGenerator(featurewise_center=True, featurewise_std_normalization=True)
for data in (train_data, validation_data):
data.fit(x_train) # 実用を考えると、x_validationでのfeaturewiseのfitは無理だと思う……。
batch_size = 100
epochs = 200
results = model.fit_generator(train_data.flow(x_train, y_train, batch_size=batch_size),
steps_per_epoch=x_train.shape[0] // batch_size,
epochs=epochs,
callbacks=[LearningRateScheduler(partial(getitem, tuple(take(epochs, concat(repeat(0.01, 1), repeat(0.1, 99), repeat(0.01, 50), repeat(0.001))))))],
validation_data=validation_data.flow(x_validation, y_validation, batch_size=batch_size),
validation_steps=x_validation.shape[0] // batch_size)
with open('./results/history.pickle', 'wb') as f:
pickle.dump(results.history, f)
save_model(model, './results/model.h5')
del model
def draw_wave(self, wave):
min_ys, max_ys = zip(
*map(juxt(np.min, np.max), np.array_split(wave * 128 + 128, 256)))
for object_id in self.wave_canvas.find_all():
self.wave_canvas.delete(object_id)
for x, min_y, max_y in zip(count(), min_ys, max_ys):
self.wave_canvas.create_line(x, min_y, x, max_y)
def ljuxt(*fs):
return rcompose(juxt(*fs), list)
def ljuxt(*fs): # Kerasはジェネレーターを引数に取るのを嫌がるみたい、かつ、funcyはPython3だと積極的にジェネレーターを使うみたいなので、リストを返すjuxtを作りました。
return rcompose(juxt(*fs), list)
def get_min_max_coordinate():
coordinates = tuple(mapcat(lambda shape: (shape.bb.left, shape.bb.top, shape.bb.right, shape.bb.bottom), filter(lambda shape: shape.body.body_type == pymunk.Body.DYNAMIC, space.shapes)))
return juxt(max, min)(coordinates)
def notnone_fn(*funcs):
return lambda val: first(filter(notnone, juxt(*funcs)(val)))
def main():
import os
with tf.device("/cpu:0"):
(x_train, y_train), (x_validation, y_validation) = load_data()
batch_size = 32
epochs = 200
input_shape = Input(shape=x_train.shape[1:])
model_file = './results/model.h5'
if os.path.exists(model_file):
model = load_model(model_file)
# with tf.device("/cpu:0"):
# validation_data = ImageDataGenerator(featurewise_center=True, featurewise_std_normalization=True)
else:
model = Model(*juxt(identity, computational_graph(y_train.shape[1]))(
input_shape))
model.compile(loss='categorical_crossentropy',
optimizer=SGD(momentum=0.9),
metrics=['accuracy'])
with tf.device("/cpu:0"):
train_data = ImageDataGenerator(featurewise_center=True,
featurewise_std_normalization=True,
width_shift_range=0.125,
height_shift_range=0.125,
horizontal_flip=True)
validation_data = ImageDataGenerator(
featurewise_center=True, featurewise_std_normalization=True)
for data in (train_data, validation_data):
data.fit(
x_train) # 実用を考えると、x_validationでのfeaturewiseのfitは無理だと思う……。
results = model.fit_generator(
train_data.flow(x_train, y_train, batch_size=batch_size),
steps_per_epoch=x_train.shape[0] // batch_size,
epochs=epochs,
callbacks=[
LearningRateScheduler(
partial(
getitem,
tuple(
take(
epochs,
concat(repeat(0.01, 1), repeat(0.1, 99),
repeat(0.01, 50), repeat(0.001))))))
],
validation_data=validation_data.flow(x_validation,
y_validation,
batch_size=batch_size),
validation_steps=x_validation.shape[0] // batch_size)
with open('./results/history.pickle', 'wb') as f:
pickle.dump(results.history, f)
save_model(model, model_file)
try:
with tf.device("/cpu:0"):
# model.summary()
# print("=== AFTER POPPING THE LAST ===")
model.layers.pop()
# model.summary()
# generate_confusion_matrix(model, x_validation, y_validation, batch_size)
# plot_model(model, to_file='./results/model.png')
except Exception as ex:
print("plot_model failed with error:", repr(ex), "\nMoving on...")
siamese(input_shape, model)
def __ljuxt(self, *fs): return rcompose(juxt(*fs), list)
