def demo_ori():
model_file = "../test_res18.prototxt"
weights_file = "../weights_resize_res18.caffemodel"
data_file = "../test_pad_30_resize_112.txt"
img_root = "/home_1/data/caffe/DTY_Side"
cfg_file = "../test_yxx_resize.yml"
pkl_dir = "../pkl_feature"
batch_size = 512
cfg_from_file(cfg_file)
net = load_net_(model_file,
weights_file=weights_file,
GPU_index=3,
batch_size=batch_size,
forward_type="test")
img_file_list = open(data_file, "r").readlines()
img_list = img_file_list[0:batch_size]
input_data, input_label, input_data_name = load_data_batch_with_label(
net, img_root, img_list)
feature_v = get_feature_information(net, model_file, input_data)
# print feature_v["prob"]
if not os.path.exists(pkl_dir):
os.makedirs(pkl_dir)
save_name = os.path.join(pkl_dir, "feature_res_{}.pkl".format(batch_size))
pickle.dump(feature_v, open(save_name, 'wb'))
def test_net_cls(args):
if args.cfg_file is not None:
cfg_from_file(args.cfg_file)
net = load_net_(args.deploy_file,
args.weights_file,
args.gpu_id,
batch_size=-1,
forward_type="test")
img_list = open(args.img_list_file, "r").readlines()
input_layer_name = net._layer_names[net._inputs[0]]
batch_size, c, h, w = net.blobs[input_layer_name].shape
epoch_num = int(len(img_list) / batch_size)
predict_res = []
for epoch_i in range(epoch_num):
if epoch_i % 10 == 0:
print "TODO: {}% ".format(1.0 * epoch_i / epoch_num * 100)
batch_img_list = img_list[epoch_i * batch_size:(epoch_i + 1) *
batch_size]
input_data, input_label, input_data_name = load_data_batch_with_label(
net, args.img_root_path, batch_img_list)
single_batch_res = predict_with_label(net, input_data, input_label,
input_data_name)
predict_res.extend(single_batch_res)
pkl_dir = os.path.join(args.save_root_path, "res_pkl")
if not os.path.exists(pkl_dir):
os.makedirs(pkl_dir)
save_name = os.path.join(
pkl_dir, "{}_res.pkl".format(
os.path.splitext(os.path.basename(args.weights_file))[0]))
pickle.dump(predict_res, open(save_name, 'wb'))
def __init__(self, P_parameters):
# get ori net weights
self.P_parameters = P_parameters
self.ori_net = load_net_(self.P_parameters.model_file,
weights_file=self.P_parameters.weights_file,
GPU_index=self.P_parameters.GPU_index,
batch_size=-1,
forward_type="test")
self.Net_composition = Net_composition_(self.P_parameters.model_file)
self.pruned_layers_set = set()
self.fixed_layers_set = set()
def create_new_net(self):
self.new_net = load_net_(self.P_parameters.new_model_file,
weights_file=None,
GPU_index=self.P_parameters.GPU_index,
batch_size=-1,
forward_type="test")
def my_test():
args = parse_args()
print(args)
args.cfg_file = "../test_yxx_resize.yml"
args.gpu_id = 3
args.img_list_file = "../test_pad_30_resize_112.txt"
args.img_root_path = "/home_1/data/caffe/DTY_Side"
args.task_root = "../model_2"
args.model_file = "test_res18"
args.model_file_train = "train_res18"
args.weights_file = "weights_resize_res18"
args.solver_file = "solver_res18"
batch_size = 128
cfg_from_file(args.cfg_file)
P_parameters = Prune_parameters_(args)
net_com = Net_composition_(P_parameters.model_file)
weights_graph = net_com.weights_layers_graph
pruning_layer_dict = {"conv1": 1}
BN_pruning_dict = {"bn_conv1": [i for i in range(63)]}
Prune_ = Prune_tools(P_parameters)
Prune_.update_model_file(pruning_layer_dict)
Prune_.create_new_net()
Weights_pruning_dict = get_all_keep_index(BN_pruning_dict,
pruning_layer_dict,
weights_graph)
Prune_.update_new_net(Weights_pruning_dict)
Prune_.save_new_weights()
model_ori = "/home_1/code/caffe_test/compress/weights_pruning/20190529/model_2/test_res18.prototxt"
weights_ori = "/home_1/code/caffe_test/compress/weights_pruning/20190529/model_2/weights_resize_res18.caffemodel"
net_scale = load_net_(model_ori,
weights_file=weights_ori,
GPU_index=3,
batch_size=batch_size,
forward_type="test")
net_scale.params["scale_conv1"][0].data[-1] = 0
net_scale.params["scale_conv1"][1].data[-1] = 0
net_ori = load_net_(model_ori,
weights_file=weights_ori,
GPU_index=3,
batch_size=batch_size,
forward_type="test")
model_pruning = "/home_1/code/caffe_test/compress/weights_pruning/20190529/model_2/test_res18_0.prototxt"
weights_pruning = "/home_1/code/caffe_test/compress/weights_pruning/20190529/model_2/weights_resize_res18_0.caffemodel"
net_pruning = load_net_(model_pruning,
weights_file=weights_pruning,
GPU_index=3,
batch_size=batch_size,
forward_type="test")
img_file_list = open(args.img_list_file, "r").readlines()
img_list = img_file_list[0:batch_size]
input_data, input_label, input_data_name = load_data_batch_with_label(
net_pruning, args.img_root_path, img_list)
out_put_pruning, net_pruning = single_batch_forward(
net_pruning, input_data)
out_put_scale, net_scale = single_batch_forward(net_scale, input_data)
out_put_ori, net_ori = single_batch_forward(net_ori, input_data)
# print "======================="
# print "pruning: "
# print out_put_pruning
# print "-----------------------"
# print "scale: "
# print out_put_scale
# print "-----------------------"
for key_ in out_put_pruning.keys():
for i in range(len(out_put_pruning[key_])):
print "------------------------"
print "scale: "
print out_put_scale[key_][i]
print "pruning: "
print out_put_pruning[key_][i]
print "ori: "
print out_put_ori[key_][i]
if __name__ == '__main__':
model_file = "../test_res18.prototxt"
weights_file = "../weights_resize_res18.caffemodel"
data_file = "../test_pad_30_resize_112.txt"
img_root = "/home_1/data/caffe/DTY_Side"
cfg_file = "../test_yxx_resize.yml"
pkl_dir = "../pkl_feature"
batch_size = 512
cfg_from_file(cfg_file)
net = load_net_(model_file,
weights_file=weights_file,
GPU_index=3,
batch_size=batch_size,
forward_type="test")
img_file_list = open(data_file, "r").readlines()
random.shuffle(img_file_list)
img_list = img_file_list[0:batch_size]
input_data, input_label, input_data_name = load_data_batch_with_label(
net, img_root, img_list)
print input_data_name
print input_label
feature_v = get_pruning_prob_each(net, model_file, input_data)
if not os.path.exists(pkl_dir):