diff --git a/train.py b/train.py
index 454d6a9..79efe62 100644
--- a/train.py
+++ b/train.py
@@ -1,6 +1,7 @@
 import config
 import trainer_step
 
+
 def main():
     cfg = config.Config(filenamequeue="../data/dataset/layout_1205.tfrecords")
     t = trainer_step.Trainer(cfg)
diff --git a/trainer_step.py b/trainer_step.py
index 1f80a71..5fbe173 100644
--- a/trainer_step.py
+++ b/trainer_step.py
@@ -11,7 +11,7 @@
 tf.compat.v1.disable_eager_execution()
 # slim = tf.contrib.slim
 
-DATA_DIR = os.path.join('..','data')
+DATA_DIR = os.path.join('..')
 class Trainer(object):
     def __init__(self, config):
         filenamequeue = tf.compat.v1.train.string_input_producer([config.filenamequeue])
@@ -58,7 +58,8 @@ def _build_model(self, filenamequeue, config):
         randomz = tf.random.normal([config.batch_size, config.z_dim])
         # testing case
         testLayout, testImgfea, testSemvec, testTexfea = self.inputs(config)
-        randomz_val = np.load(os.path.join(DATA_DIR,'sample','noiseVector_128.npy'))
+        # randomz_val = np.load(os.path.join(DATA_DIR,'/drive/MyDrive/LayoutNetme/sample','noiseVector_128.npy'))
+        randomz_val = np.load('../drive/MyDrive/LayoutNetme/sample/noiseVector_128.npy')
         testSemvec, _ = model.embeddingSemvec(testSemvec, training_, reuse=True)
         testImgfea, _ = model.embeddingImg(testImgfea, training_, reuse=True)
         testTexfea, _ = model.embeddingTex(testTexfea, training_, reuse=True)
@@ -223,8 +224,8 @@ def sample(self):
         return (inputdata + 1) / 2.0, (gen + 1) / 2.0, fea
 
     def testing(self):
-        new_saver = tf.compat.v1.train.import_meta_graph('./log/layoutNet-100.meta')
-        new_saver.restore(self.sess, './log/layoutNet-100')
+        new_saver = tf.compat.v1.train.import_meta_graph('/layoutNet-100.meta')
+        new_saver.restore(self.sess, '/layoutNet-100')
         gen, fea = self.sess.run([self.model["Gtest"],
                                   self.model["Etest"]],
                                   feed_dict={self.model["is_training"]: False})
@@ -232,12 +233,12 @@ def testing(self):
         return (gen + 1) / 2.0, fea
 
     def inputs(self, config):
-        layoutpath = DATA_DIR + '/sample/layout/'
-        imgfeapath = DATA_DIR + '/sample/visfea/'
-        texfeapath = DATA_DIR + '/sample/texfea/'
-        semvecpath = DATA_DIR + '/sample/semvec/'
+        layoutpath = DATA_DIR + '/drive/MyDrive/LayoutNetme/validation/layout/'
+        imgfeapath = DATA_DIR + '/drive/MyDrive/LayoutNetme/validation/visfea/'
+        texfeapath = DATA_DIR + '/drive/MyDrive/LayoutNetme/validation/texfea/'
+        semvecpath = DATA_DIR + '/drive/MyDrive/LayoutNetme/validation/semvec/'
 
-        f = open((DATA_DIR + '/sample/imgSel_128.txt'), 'r')
+        f = open((DATA_DIR + '/drive/MyDrive/LayoutNetme/validation/imgSel_60.txt'), 'r')
         name = f.read()
         namelist = name.split()
         n_samples = len(namelist)