def main(batch_sizes, crop_size, random_move, learning_rate,
segmentation_task_ratio, weight_decay, save_folder, epochs):
'''
:param batch_sizes: the number of examples of each class in a single batch
:param crop_size: the input size
:param random_move: the random move in data augmentation
:param learning_rate: learning rate of the optimizer
:param segmentation_task_ratio: the weight of segmentation loss in total loss
:param weight_decay: l2 weight decay
:param save_folder: where to save the snapshots, tensorflow logs, etc.
:param epochs: how many epochs to run
:return:
'''
batch_size = sum(batch_sizes)
train_dataset = ClfDataset(crop_size=crop_size, subset=[0, 1, 2], move=None,
define_label=lambda l: [l[0] ,l[1]])
val_dataset = ClfDataset(crop_size=crop_size, subset=[3], move=None,
define_label=lambda l: [l[0] ,l[1]])
train_loader = get_balanced_loader(train_dataset, batch_sizes=batch_sizes)
val_loader = get_loader(val_dataset, batch_size=batch_size)
model = densesharp.get_compiled(output_size=2,
optimizer=Adam(lr=learning_rate),
loss='categorical_crossentropy',
metrics=['accuracy'],
weight_decay=weight_decay)
checkpointer = ModelCheckpoint(filepath='tmp/%s/weights.{epoch:02d}.h5' % save_folder, verbose=1,
period=1, save_weights_only=True)
best_keeper = ModelCheckpoint(filepath='tmp/%s/best.h5' % save_folder, verbose=1, save_weights_only=True,
monitor='val_clf_acc', save_best_only=True, period=1, mode='max')
csv_logger = CSVLogger('tmp/%s/training.csv' % save_folder)
tensorboard = TensorBoard(log_dir='tmp/%s/logs/' % save_folder)
early_stopping = EarlyStopping(monitor='val_clf_acc', min_delta=0, mode='max',
patience=30, verbose=1)
lr_reducer = ReduceLROnPlateau(monitor='val_loss', factor=0.334, patience=10,
verbose=1, mode='min', epsilon=1.e-5, cooldown=2, min_lr=0)
#model.fit_generator(generator=train_loader, steps_per_epoch=len(train_dataset), max_queue_size=500, workers=1,
# validation_data=val_loader, epochs=epochs, validation_steps=len(val_dataset))
model.fit_generator(generator=train_loader, steps_per_epoch=len(train_dataset), max_queue_size=500, workers=1,
validation_data=val_loader, epochs=epochs, validation_steps=len(val_dataset),
callbacks=[checkpointer, early_stopping, best_keeper, lr_reducer, csv_logger])
#save model
with open('cnns1225.pickle','wb') as file:
pickle.dump(model,file)
test=get_x_test()
print(np.shape(test))
y=model.predict(test)
for i in range(117):
print(y[i,1])
def main(batch_sizes, crop_size, random_move, learning_rate,
segmentation_task_ratio, weight_decay, save_folder, epochs):
'''
:param batch_sizes: the number of examples of each class in a single batch
:param crop_size: the input size
:param random_move: the random move in data augmentation
:param learning_rate: learning rate of the optimizer
:param segmentation_task_ratio: the weight of segmentation loss in total loss
:param weight_decay: l2 weight decay
:param save_folder: where to save the snapshots, tensorflow logs, etc.
:param epochs: how many epochs to run
:return:
'''
batch_size = sum(batch_sizes)
train_dataset = ClfSegDataset(crop_size=crop_size, subset=[0, 1, 2, 3], move=random_move,
define_label=lambda l: [l[0] + l[1], l[2], l[3]])
val_dataset = ClfSegDataset(crop_size=crop_size, subset=[4], move=None,
define_label=lambda l: [l[0] + l[1], l[2], l[3]])
train_loader = get_balanced_loader(train_dataset, batch_sizes=batch_sizes)
val_loader = get_loader(val_dataset, batch_size=batch_size)
model = densesharp.get_compiled(output_size=3,
optimizer=Adam(lr=learning_rate),
loss={"clf": 'categorical_crossentropy',
"seg": losses.DiceLoss()},
metrics={'clf': ['accuracy', metrics.precision, metrics.recall, metrics.fmeasure,
metrics.invasion_acc, metrics.invasion_fmeasure,
metrics.invasion_precision, metrics.invasion_recall,
metrics.ia_acc, metrics.ia_fmeasure,
metrics.ia_precision, metrics.ia_recall],
'seg': [metrics.precision, metrics.recall, metrics.fmeasure]},
loss_weights={"clf": 1., "seg": segmentation_task_ratio},
weight_decay=weight_decay)
checkpointer = ModelCheckpoint(filepath='tmp/%s/weights.{epoch:02d}.h5' % save_folder, verbose=1,
period=1, save_weights_only=True)
best_keeper = ModelCheckpoint(filepath='tmp/%s/best.h5' % save_folder, verbose=1, save_weights_only=True,
monitor='val_clf_acc', save_best_only=True, period=1, mode='max')
csv_logger = CSVLogger('tmp/%s/training.csv' % save_folder)
tensorboard = TensorBoard(log_dir='tmp/%s/logs/' % save_folder)
early_stopping = EarlyStopping(monitor='val_clf_acc', min_delta=0, mode='max',
patience=30, verbose=1)
lr_reducer = ReduceLROnPlateau(monitor='val_loss', factor=0.334, patience=10,
verbose=1, mode='min', epsilon=1.e-5, cooldown=2, min_lr=0)
model.fit_generator(generator=train_loader, steps_per_epoch=len(train_dataset), max_queue_size=500, workers=1,
validation_data=val_loader, epochs=epochs, validation_steps=len(val_dataset),
callbacks=[checkpointer, early_stopping, best_keeper, lr_reducer, csv_logger, tensorboard])
def main():
model = densesharp.get_compiled(output_size=2,
optimizer=Adam(lr=1.e-4),
loss='categorical_crossentropy',
metrics=['accuracy'],
weight_decay=0)
model.load_weights('result/p4.h5')
with open('result/p4.pickle', 'wb') as file:
pickle.dump(model, file)
test = get_x_test()
print(np.shape(test))
y = model.predict(test)
for i in range(117):
print(y[i, 1])
del X_test
X_test_new = X_test_new.reshape(X_test_new.shape[0], 32, 32, 32,
1) #将测试数据集整合成5d张量
print(X_test_new.shape)
model = densesharp.get_compiled(
output_size=2,
optimizer=Adam(lr=1.e-4),
loss={
"clf": 'categorical_crossentropy',
"seg": losses.DiceLoss()
},
metrics={
'clf': [
'accuracy', metrics.precision, metrics.recall, metrics.fmeasure,
metrics.invasion_acc, metrics.invasion_fmeasure,
metrics.invasion_precision, metrics.invasion_recall,
metrics.ia_acc, metrics.ia_fmeasure, metrics.ia_precision,
metrics.ia_recall
],
'seg': [metrics.precision, metrics.recall, metrics.fmeasure]
},
loss_weights={
"clf": 1.,
"seg": 0.2
},
weight_decay=0)
filepath = "best_weight_sharp.h5"
save_folder = 'test'
checkpointer = ModelCheckpoint(filepath='tmp/%s/weights.{epoch:02d}.h5' %