diff --git a/AirShower_WGAN/ShowerGAN.py b/AirShower_WGAN/ShowerGAN.py
index bb7d36fd7a6e4bda0f27483d94d1a9dab02650f2..c40cfb528302e0a55d428d4865b4334e2daa04c1 100644
--- a/AirShower_WGAN/ShowerGAN.py
+++ b/AirShower_WGAN/ShowerGAN.py
@@ -7,20 +7,25 @@ from functools import partial
 from keras.utils import plot_model
 import keras.backend.tensorflow_backend as KTF
 import utils
+import glob
 
 
 KTF.set_session(utils.get_session())  # Allows 2 jobs per GPU, Please do not change this during the tutorial
 log_dir = "."
 
 # basic trainings parameter
-EPOCHS = 10
+EPOCHS = 20
 GRADIENT_PENALTY_WEIGHT = 10
 BATCH_SIZE = 256
 NCR = 5
 latent_size = 512
 
 # load trainings data
-shower_maps, Energy = utils.ReadInData()
+Folder = '/net/scratch/lgeiger/refiner_paper/data/toy-data/pre_pro'
+filenames=glob.glob(Folder + "/showers_pre_pro_*")
+print filenames
+shower_maps, Energy = utils.ReadInData(filenames)
+shower_maps = shower_maps[:,:,:,1,np.newaxis]
 N = shower_maps.shape[0]
 # plot real signal patterns
 utils.plot_multiple_signalmaps(shower_maps[:, :, :, 0], log_dir=log_dir, title='Footprints', epoch='Real')
diff --git a/AirShower_WGAN/utils.py b/AirShower_WGAN/utils.py
index af15aa32ba53e334853da9f5c4907eb9fba151d2..c070a604745e8c4e531fef75cbddf44022a1733a 100644
--- a/AirShower_WGAN/utils.py
+++ b/AirShower_WGAN/utils.py
@@ -15,11 +15,26 @@ def get_session(gpu_fraction=0.40):
     return tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
 
 
-def ReadInData():
-    '''Reads in the trainings data'''
-    filenames = "/net/scratch/JGlombitza/Public/HAPWorkshop2018/data/Data.npz"
-    data = np.load(filenames)
-    return data['shower_maps'], data['Energy']
+def ReadInData(filenames):
+    N = 10000 *len(filenames)
+    a = 10000 # packagesize
+    Input2 = np.zeros(N*9*9*2).reshape(N,9,9,2)
+    Energy = np.zeros(N)
+    showeraxis = np.zeros(N*3).reshape(N,3)
+    
+    for i in range(0, len(filenames)):
+        data = np.load(filenames[i])
+        try:
+            Energy[a*i:a*(i+1)] = data['energy']
+            Input2[a*i:a*(i+1)] = data['input2'].reshape(a,9,9,2)
+        except:
+            Energy[a*i:a*(i+1)] = data['Energy']
+            Input2[a*i:a*(i+1)] = data['Input2'].reshape(a,9,9,2)
+    Input2[:,:,:,0] = Input2[:,:,:,0] - np.min(Input2[:,:,:,0], axis=(1,2))[:,np.newaxis,np.newaxis] + 1
+    Filter = Input2[:,:,:,1] == 0
+    Input2[:,:,:,0][Filter] = 0
+    return Input2, Energy
+
 
 
 def make_trainable(model, trainable):