diff --git a/README.md b/README.md
index 945587e43b3fb1271a3b49c34972f4f5f04553ea..7d30cc9360207019cd245941c6749191f732bcc0 100644
--- a/README.md
+++ b/README.md
@@ -40,7 +40,7 @@ cd candock
 python3 train.py --label 50 --input_nc 1 --dataset_dir ./datasets/simple_test --save_dir ./checkpoints/simple_test --model_name micro_multi_scale_resnet_1d --gpu_id 0 --batchsize 64 --k_fold 5
 # if you want to use cpu to train, please input --no_cuda
 ```
-* More [options](./options.py).
+* More [options](./util/options.py).
 #### Use your own data to train
 * step1: Generate signals.npy and labels.npy in the following format.
 ```python
@@ -56,4 +56,5 @@ labels = np.array([0,0,0,0,0,1,1,1,1,1])      #0->class0    1->class1
 ```bash
 python3 simple_test.py --label 50 --input_nc 1 --model_name micro_multi_scale_resnet_1d --gpu_id 0
 # if you want to use cpu to test, please input --no_cuda
-```
\ No newline at end of file
+```
+
diff --git a/confusion_mat b/docs/confusion_mat
similarity index 100%
rename from confusion_mat
rename to docs/confusion_mat
diff --git a/creatnet.py b/models/creatnet.py
similarity index 85%
rename from creatnet.py
rename to models/creatnet.py
index f0ba6f890fe9e65088154cca9913308363e0b10a..eb16fbf8ec7a468044fb0e6931534a6324ffb830 100644
--- a/creatnet.py
+++ b/models/creatnet.py
@@ -1,6 +1,8 @@
 from torch import nn
-from models import cnn_1d,densenet,dfcnn,lstm,mobilenet,resnet,resnet_1d,squeezenet
-from models import multi_scale_resnet,multi_scale_resnet_1d,micro_multi_scale_resnet_1d
+from . import cnn_1d,densenet,dfcnn,lstm,mobilenet,resnet,resnet_1d,squeezenet, \
+multi_scale_resnet,multi_scale_resnet_1d,micro_multi_scale_resnet_1d
+# from models import cnn_1d,densenet,dfcnn,lstm,mobilenet,resnet,resnet_1d,squeezenet
+# from models import multi_scale_resnet,multi_scale_resnet_1d,micro_multi_scale_resnet_1d
 
 def CreatNet(opt):
     name = opt.model_name
diff --git a/simple_test.py b/simple_test.py
index efe48793bdb140a7c7b2c7cae09eae037c49f46b..de0cf2746b09942eb2745a80e64d310b90e21123 100644
--- a/simple_test.py
+++ b/simple_test.py
@@ -3,19 +3,16 @@ import numpy as np
 import torch
 import matplotlib.pyplot as plt
 
-import util
-import transformer
-import dataloader
-from options import Options
-from creatnet import CreatNet
+from util import util,transformer,dataloader,statistics,heatmap,options
+from models import creatnet
 
 '''
 --------------------------------preload data--------------------------------
 @hypox64
 2020/04/03
 '''
-opt = Options().getparse()
-net = CreatNet(opt)
+opt = options.Options().getparse()
+net = creatnet.CreatNet(opt)
 
 #load data
 signals = np.load('./datasets/simple_test/signals.npy')
diff --git a/train.py b/train.py
index 5df13170e375213421790583ff7eed2062d9ad30..be7bd12fb03948c619e10829f5e5bda8e2a11d56 100644
--- a/train.py
+++ b/train.py
@@ -7,15 +7,10 @@ from torch import nn, optim
 import warnings
 warnings.filterwarnings("ignore")
 
-import util
-import transformer
-import dataloader
-import statistics
-import heatmap
-from creatnet import CreatNet
-from options import Options
-
-opt = Options().getparse()
+from util import util,transformer,dataloader,statistics,heatmap,options
+from models import creatnet
+
+opt = options.Options().getparse()
 torch.cuda.set_device(opt.gpu_id)
 t1 = time.time()
 
@@ -37,7 +32,7 @@ train_sequences,test_sequences = transformer.k_fold_generator(len(labels),opt.k_
 t2 = time.time()
 print('load data cost time: %.2f'% (t2-t1),'s')
 
-net=CreatNet(opt)
+net=creatnet.CreatNet(opt)
 util.writelog('network:\n'+str(net),opt,True)
 
 util.show_paramsnumber(net,opt)
diff --git a/util/__init__.py b/util/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/dataloader.py b/util/dataloader.py
similarity index 99%
rename from dataloader.py
rename to util/dataloader.py
index 6906e436db4e87150386cd16ee767953e41fc38c..6866f1611af224000d3c0c159319ccb6e7216e77 100644
--- a/dataloader.py
+++ b/util/dataloader.py
@@ -5,9 +5,10 @@ import random
 import scipy.io as sio
 import numpy as np
 
-import dsp
-import transformer
-import statistics
+from . import dsp,transformer,statistics
+# import dsp
+# import transformer
+# import statistics
 
 
 def trimdata(data,num):
diff --git a/dsp.py b/util/dsp.py
similarity index 100%
rename from dsp.py
rename to util/dsp.py
diff --git a/heatmap.py b/util/heatmap.py
similarity index 100%
rename from heatmap.py
rename to util/heatmap.py
diff --git a/options.py b/util/options.py
similarity index 98%
rename from options.py
rename to util/options.py
index 9905cf6b68c10cb774322d047d53b683adfb17de..44559ba33f1e775d65b1cfc992692c304ab0622a 100644
--- a/options.py
+++ b/util/options.py
@@ -1,7 +1,7 @@
 import argparse
 import os
 import time
-import util
+from . import util
 
 class Options():
     def __init__(self):
@@ -25,7 +25,7 @@ class Options():
         self.parser.add_argument('--continue_train', action='store_true', help='if specified, continue train')
         self.parser.add_argument('--lr', type=float, default=0.001,help='learning rate') 
         self.parser.add_argument('--batchsize', type=int, default=64,help='batchsize')
-        self.parser.add_argument('--weight_mod', type=str, default='normal',help='Choose weight mode: auto | normal')
+        self.parser.add_argument('--weight_mod', type=str, default='auto',help='Choose weight mode: auto | normal')
         self.parser.add_argument('--epochs', type=int, default=20,help='end epoch')
         self.parser.add_argument('--network_save_freq', type=int, default=5,help='the freq to save network')
         self.parser.add_argument('--k_fold', type=int, default=0,help='fold_num of k-fold.if 0 or 1,no k-fold')
diff --git a/statistics.py b/util/statistics.py
similarity index 99%
rename from statistics.py
rename to util/statistics.py
index ebabbd1cc62719a8b568ed7993f8ce7be04c5b50..11c0c2b87ac518af72030739c241fb327def3680 100644
--- a/statistics.py
+++ b/util/statistics.py
@@ -2,7 +2,7 @@ import numpy as np
 import matplotlib.pyplot as plt
 import util
 import os
-import heatmap
+from . import heatmap
 
 def label_statistics(labels):
     #for sleep label: N3->0  N2->1  N1->2  REM->3  W->4
diff --git a/transformer.py b/util/transformer.py
similarity index 99%
rename from transformer.py
rename to util/transformer.py
index 18b860214a41c36eb30e2191f7ab15119b4901ac..7098ad579f857d7849a4ebcb7c885ce66be87178 100644
--- a/transformer.py
+++ b/util/transformer.py
@@ -2,7 +2,8 @@ import os
 import random
 import numpy as np
 import torch
-import dsp
+from . import dsp
+# import dsp
 
 def trimdata(data,num):
     return data[:num*int(len(data)/num)]
diff --git a/util.py b/util/util.py
similarity index 100%
rename from util.py
rename to util/util.py