diff --git a/model.py b/model.py
index 0bc4576..48e36eb 100644
--- a/model.py
+++ b/model.py
@@ -57,14 +57,31 @@ def conv2d(x, W, stride):
 keep_prob = tf.placeholder(tf.float32)
 h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
 
+#FCL12
+W_fc12 = weight_variable([1164, 800])
+b_fc12 = bias_variable([800])
+
+h_fc12 = tf.nn.relu(tf.matmul(h_fc1_drop, W_fc12) + b_fc12)
+
+h_fc12_drop = tf.nn.dropout(h_fc12, keep_prob)
+
 #FCL 2
-W_fc2 = weight_variable([1164, 100])
+W_fc2 = weight_variable([800, 100])
 b_fc2 = bias_variable([100])
 
-h_fc2 = tf.nn.relu(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
+h_fc2 = tf.nn.relu(tf.matmul(h_fc12_drop, W_fc2) + b_fc2)
 
 h_fc2_drop = tf.nn.dropout(h_fc2, keep_prob)
 
+
+#FCL 2
+#W_fc2 = weight_variable([1164, 100])
+#b_fc2 = bias_variable([100])
+
+#h_fc2 = tf.nn.relu(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
+
+#h_fc2_drop = tf.nn.dropout(h_fc2, keep_prob)
+
 #FCL 3
 W_fc3 = weight_variable([100, 50])
 b_fc3 = bias_variable([50])
@@ -85,4 +102,4 @@ def conv2d(x, W, stride):
 W_fc5 = weight_variable([10, 1])
 b_fc5 = bias_variable([1])
 
-y = tf.mul(tf.atan(tf.matmul(h_fc4_drop, W_fc5) + b_fc5), 2) #scale the atan output
+y = tf.multiply(tf.atan(tf.matmul(h_fc4_drop, W_fc5) + b_fc5), 2) #scale the atan output
diff --git a/run.py b/run.py
index fa94c3f..537a2ea 100644
--- a/run.py
+++ b/run.py
@@ -18,7 +18,7 @@
     ret, frame = cap.read()
     image = scipy.misc.imresize(frame, [66, 200]) / 255.0
     degrees = model.y.eval(feed_dict={model.x: [image], model.keep_prob: 1.0})[0][0] * 180 / scipy.pi
-    call("clear")
+    #call("clear")
     print("Predicted steering angle: " + str(degrees) + " degrees")
     cv2.imshow('frame', frame)
     #make smooth angle transitions by turning the steering wheel based on the difference of the current angle
diff --git a/run_dataset.py b/run_dataset.py
index 1c1a87d..86e01a1 100644
--- a/run_dataset.py
+++ b/run_dataset.py
@@ -9,6 +9,7 @@
 saver.restore(sess, "save/model.ckpt")
 
 img = cv2.imread('steering_wheel_image.jpg',0)
+#img = cv2.imread('me0.jpg',0)
 rows,cols = img.shape
 
 smoothed_angle = 0
@@ -18,7 +19,7 @@
     full_image = scipy.misc.imread("driving_dataset/" + str(i) + ".jpg", mode="RGB")
     image = scipy.misc.imresize(full_image[-150:], [66, 200]) / 255.0
     degrees = model.y.eval(feed_dict={model.x: [image], model.keep_prob: 1.0})[0][0] * 180.0 / scipy.pi
-    call("clear")
+    #call("clr")
     print("Predicted steering angle: " + str(degrees) + " degrees")
     cv2.imshow("frame", cv2.cvtColor(full_image, cv2.COLOR_RGB2BGR))
     #make smooth angle transitions by turning the steering wheel based on the difference of the current angle
diff --git a/save/model.ckpt.meta b/save/model.ckpt.meta
index 8afe34b..9d902bd 100644
Binary files a/save/model.ckpt.meta and b/save/model.ckpt.meta differ
diff --git a/train.py b/train.py
index 4d6a6a2..de52a17 100644
--- a/train.py
+++ b/train.py
@@ -11,23 +11,26 @@
 
 train_vars = tf.trainable_variables()
 
-loss = tf.reduce_mean(tf.square(tf.sub(model.y_, model.y))) + tf.add_n([tf.nn.l2_loss(v) for v in train_vars]) * L2NormConst
+loss = tf.reduce_mean(tf.square(tf.subtract(model.y_, model.y))) + tf.add_n([tf.nn.l2_loss(v) for v in train_vars]) * L2NormConst
 train_step = tf.train.AdamOptimizer(1e-4).minimize(loss)
-sess.run(tf.initialize_all_variables())
+#sess.run(tf.initialize_all_variables())
+sess.run(tf.global_variables_initializer())
 
 # create a summary to monitor cost tensor
-tf.scalar_summary("loss", loss)
+tf.summary.scalar("loss", loss)
 # merge all summaries into a single op
-merged_summary_op = tf.merge_all_summaries()
+#merged_summary_op = tf.merge_all_summaries()
+merged_summary_op = tf.summary.merge_all()
 
 saver = tf.train.Saver()
 
 # op to write logs to Tensorboard
 logs_path = './logs'
-summary_writer = tf.train.SummaryWriter(logs_path, graph=tf.get_default_graph())
+#summary_writer = tf.train.SummaryWriter(logs_path, graph=tf.get_default_graph())
+summary_writer = tf.summary.FileWriter(logs_path, graph=tf.get_default_graph())
 
-epochs = 30
-batch_size = 100
+epochs = 3
+batch_size = 10
 
 # train over the dataset about 30 times
 for epoch in range(epochs):