def generate_single_svm_train(train_file):
save_path = train_file.rsplit('.', 1)[0].strip()
if not os.path.exists(save_path):
os.makedirs(save_path)
if len(os.listdir(save_path)) == 0:
print("reading %s's svm dataset" % train_file.split('\\')[-1])
# 避免正负样本数量相差过大，分类器偏向负，所以<0.1的为负，大于0.6的为正，其他作为测试样本
# 要保证第一次分类后有一些测试样本被判为正样本，方便后面难负例挖掘
prep.load_train_proposals(train_file, 2, save_path, threshold=0.1, is_svm=True, save=True)
print("restoring svm dataset")
images, labels, rects = prep.load_from_npy(save_path, is_svm=True)
return images, labels, rects
def generate_single_svm_train(train_file):
save_path = train_file.rsplit('.', 1)[0].strip()
if len(os.listdir(save_path)) == 0:
print("reading %s's svm dataset" % train_file.split('\\')[-1])
prep.load_train_proposals(train_file,
2,
save_path,
threshold=0.3,
is_svm=True,
save=True)
print("restoring svm dataset")
images, labels = prep.load_from_npy(save_path)
return images, labels
def generate_single_svm_train(one_class_train_file):
trainfile = one_class_train_file
savepath = one_class_train_file.replace('txt', 'pkl')
images = []
Y = []
if os.path.isfile(savepath):
print("restoring svm dataset " + savepath)
images, Y = prep.load_from_pkl(savepath)
else:
print("loading svm dataset " + savepath)
images, Y = prep.load_train_proposals(trainfile, 2, threshold=0.3, svm=True, save=True, save_path=savepath)
return images, Y
def generate_single_svm_train(one_class_train_file):
trainfile = one_class_train_file
savepath = one_class_train_file.replace('txt', 'pkl')
images = []
Y = []
if os.path.isfile(savepath):
print("restoring svm dataset " + savepath)
images, Y = prep.load_from_pkl(savepath)
else:
print("loading svm dataset " + savepath)
images, Y = prep.load_train_proposals(trainfile,
2,
threshold=0.3,
svm=True,
save=True,
save_path=savepath)
return images, Y
n_epoch=20,
validation_set=0.1,
shuffle=True,
show_metric=True,
batch_size=64,
snapshot_step=200,
snapshot_epoch=False,
run_id='alexnet_rcnnflowers2'
) # epoch = 1000 Start training (apply gradient descent algorithm)
# Save the model
model.save('fine_tune_model_save.model')
if __name__ == '__main__':
if os.path.isfile('dataset.pkl'):
print("Loading Data")
X, Y = prep.load_from_pkl('dataset.pkl')
else:
print("Reading Data")
X, Y = prep.load_train_proposals('refine_list.txt', 2, save=True)
print("DONE")
restore = False
if os.path.isfile('fine_tune_model_save.model'):
restore = True
print("Continue training")
net = create_alexnet(3, restore)
fine_tune_Alexnet(net, X, Y)
n_epoch=1,
validation_set=0.1,
shuffle=True,
show_metric=True,
batch_size=64,
snapshot_step=200,
snapshot_epoch=False,
run_id='alexnet_rcnnflowers2')
# Save the model
model.save(fine_tune_model_path)
if __name__ == '__main__':
data_set = config.FINE_TUNE_DATA
if len(os.listdir(config.FINE_TUNE_DATA)) == 0:
print("Reading Data")
prep.load_train_proposals(config.FINE_TUNE_LIST,
2,
save=True,
save_path=data_set)
print("Loading Data")
X, Y = prep.load_from_npy(data_set)
restore = False
if os.path.isfile(config.FINE_TUNE_MODEL_PATH + '.index'):
restore = True
print("Continue fine-tune")
# three classes include background
net = create_alexnet(config.FINE_TUNE_CLASS, restore=restore)
fine_tune_Alexnet(net, X, Y, config.SAVE_MODEL_PATH,
config.FINE_TUNE_MODEL_PATH)
print("Loading the fine tuned model")
model.load('fine_tune_model_save.model')
elif os.path.isfile('model_save.model'):
print("Loading the alexnet")
model.load('model_save.model')
else:
print("No file to load, error")
return False
model.fit(X, Y, n_epoch=10, validation_set=0.1, shuffle=True,
show_metric=True, batch_size=64, snapshot_step=200,
snapshot_epoch=False, run_id='alexnet_rcnnflowers2') # epoch = 1000
# Save the model
model.save('fine_tune_model_save.model')
if __name__ == '__main__':
if os.path.isfile('dataset.pkl'):
print("Loading Data")
X, Y = prep.load_from_pkl('dataset.pkl')
else:
print("Reading Data")
X, Y = prep.load_train_proposals('refine_list.txt', 2, save=True)
print("DONE")
restore = False
if os.path.isfile('fine_tune_model_save.model'):
restore = True
print("Continue training")
net = create_alexnet(3, restore)
fine_tune_Alexnet(net,X,Y)