diff --git a/1-grid-world/6-deep-sarsa/deep_sarsa_agent.py b/1-grid-world/6-deep-sarsa/deep_sarsa_agent.py
index 1af7fda..3384f5d 100755
--- a/1-grid-world/6-deep-sarsa/deep_sarsa_agent.py
+++ b/1-grid-world/6-deep-sarsa/deep_sarsa_agent.py
@@ -3,9 +3,13 @@
 import random
 import numpy as np
 from environment import Env
-from keras.layers import Dense
-from keras.optimizers import Adam
-from keras.models import Sequential
+#from keras.layers import Dense
+#from keras.optimizers import Adam
+#from keras.models import Sequential
+
+from tensorflow.keras.layers import Dense
+from tensorflow.keras.optimizers import Adam
+from tensorflow.keras.models import Sequential
 
 EPISODES = 1000
 
diff --git a/1-grid-world/6-deep-sarsa/deep_sarsa_agent_v2.py b/1-grid-world/6-deep-sarsa/deep_sarsa_agent_v2.py
new file mode 100755
index 0000000..85b1216
--- /dev/null
+++ b/1-grid-world/6-deep-sarsa/deep_sarsa_agent_v2.py
@@ -0,0 +1,156 @@
+from __future__ import absolute_import, division, print_function, unicode_literals
+import copy
+import pylab
+import random
+import numpy as np
+from environment import Env
+
+import tensorflow as tf
+tf.config.experimental_run_functions_eagerly(True)
+from tensorflow.keras import backend
+import tensorflow.keras.layers as layers
+import tensorflow.keras.losses as losses
+import tensorflow.keras.models as models
+import tensorflow.keras.optimizers as optim
+
+EPISODES = 1000
+
+
+class DeepSARSA(tf.keras.Model):
+    def __init__(self, state_size, action_size, learning_rate,
+                 epsilon, epsilon_min, epsilon_decay, discount_factor):
+        super(DeepSARSA, self).__init__()
+
+        self.state_size = state_size
+        self.action_size = action_size
+        self.learning_rate = learning_rate
+        self.epsilon = epsilon
+        self.epsilon_min = epsilon_min
+        self.epsilon_decay = epsilon_decay
+        self.discount_factor = discount_factor
+
+        self.model = models.Sequential()
+        self.model.add(layers.Dense(30, activation='relu'))
+        self.model.add(layers.Dense(30, activation='relu'))
+        self.model.add(layers.Dense(self.action_size, activation='linear'))
+
+        self.compile(loss='mse', optimizer=optim.Adam(lr=self.learning_rate))
+        self.build(input_shape=(1, self.state_size))
+        self.summary()
+
+    def call(self, inputs):
+        x = tf.convert_to_tensor(inputs)
+
+        q_values = self.model(x)
+        #x = self.hidden1(x)
+        #x = self.hidden2(x)
+
+        #q_values = self.linear(x)
+
+        return q_values
+
+    def get_action(self, state):
+        if np.random.rand() <= self.epsilon:
+            return random.randrange(self.action_size)
+        else:
+            state = np.float32(state)
+            q_values = self.call(state)
+            return np.argmax(q_values[0])
+
+    def update(self, state, action, reward, next_state, next_action, done):
+
+        if self.epsilon > self.epsilon_min:
+            self.epsilon *= self.epsilon_decay
+
+        state = np.float32(state)
+        next_state = np.float32(next_state)
+
+        target = self.predict(state)[0]
+        # 살사의 큐함수 업데이트 식
+        if done:
+            target[action] = reward
+        else:
+            next_q = self.predict(next_state)[0]
+            # tf.assign(target[action], [(reward + self.discount_factor * next_q[next_action])])
+            target[action] = (reward + self.discount_factor * next_q[next_action])
+
+        # 출력 값 reshape
+        target = np.reshape(target, [1, 5])
+        # 인공신경망 업데이트
+        self.fit(state, target, epochs=1, verbose=0)
+
+
+# 그리드월드 예제에서의 딥살사 에이전트
+class Trainer:
+    def __init__(self, env, load_model=None):
+        self.env = env
+        # 에이전트가 가능한 모든 행동 정의
+        self.action_space = [0, 1, 2, 3, 4]
+        # 상태의 크기와 행동의 크기 정의
+        self.action_size = len(self.action_space)
+        self.state_size = 15
+        self.discount_factor = 0.99
+        self.learning_rate = 0.001
+
+        self.epsilon = 1.  # exploration
+        self.epsilon_decay = .9999
+        self.epsilon_min = 0.01
+        self.agent = DeepSARSA(state_size=self.state_size,
+                               action_size=self.action_size,
+                               learning_rate=self.learning_rate,
+                               epsilon=self.epsilon,
+                               epsilon_min=0.01,
+                               epsilon_decay=.9999,
+                               discount_factor=0.99)
+
+        self.global_step = 0
+        self.scores, self.episodes = [], []
+
+        if load_model is not None:
+            self.epsilon = 0.05
+            self.agent = models.load_model('./save_model/deep_sarsa_pretrained.h5')
+
+    def train(self):
+
+        for e in range(EPISODES):
+            done = False
+            score = 0
+            state = self.env.reset()
+            state = np.reshape(state, [1, 15])
+
+            while not done:
+
+                self.global_step += 1
+
+                action = self.agent.get_action(state)
+
+                next_state, reward, done = env.step(action)
+                next_state = np.reshape(next_state, [1, 15])
+                next_action = self.agent.get_action(next_state)
+
+                self.agent.update(state, action, reward,
+                                  next_state, next_action, done)
+
+                state = next_state
+                score += reward
+
+                state = copy.deepcopy(next_state)
+
+                if done:
+                    self.scores.append(score)
+                    self.episodes.append(e)
+                    pylab.plot(self.episodes, self.scores, 'b')
+                    pylab.savefig("./save_graph/deep_sarsa_plot.png")
+
+                    print("Episodes : {:04d}, Score: {:.2f}, Global_step : {:03d}, Epsilon : {:.4f}"
+                          .format(e, score, self.global_step, self.agent.epsilon))
+
+                if e % 100 == 0:
+                    self.agent.model.save('./save_model/deep_sarsa_pretrained.h5')
+
+
+if __name__ == "__main__":
+    # 환경과 에이전트 생성
+    env = Env()
+    trainer = Trainer(env)
+    trainer.train()
diff --git a/1-grid-world/6-deep-sarsa/save_graph/deep_sarsa_.png b/1-grid-world/6-deep-sarsa/save_graph/deep_sarsa_.png
new file mode 100644
index 0000000..978ff12
Binary files /dev/null and b/1-grid-world/6-deep-sarsa/save_graph/deep_sarsa_.png differ
diff --git a/1-grid-world/6-deep-sarsa/save_model/deep_sarsa.h5 b/1-grid-world/6-deep-sarsa/save_model/deep_sarsa.h5
new file mode 100644
index 0000000..30ca91a
Binary files /dev/null and b/1-grid-world/6-deep-sarsa/save_model/deep_sarsa.h5 differ
diff --git a/img/circle.png b/img/circle.png
new file mode 100644
index 0000000..7aeacd3
Binary files /dev/null and b/img/circle.png differ
diff --git a/img/down.png b/img/down.png
new file mode 100644
index 0000000..cd94e13
Binary files /dev/null and b/img/down.png differ
diff --git a/img/left.png b/img/left.png
new file mode 100644
index 0000000..079c57b
Binary files /dev/null and b/img/left.png differ
diff --git a/img/rectangle.png b/img/rectangle.png
new file mode 100644
index 0000000..b7cea07
Binary files /dev/null and b/img/rectangle.png differ
diff --git a/img/right.png b/img/right.png
new file mode 100644
index 0000000..cbe1b1b
Binary files /dev/null and b/img/right.png differ
diff --git a/img/triangle.png b/img/triangle.png
new file mode 100644
index 0000000..1cd9db0
Binary files /dev/null and b/img/triangle.png differ
diff --git a/img/up.png b/img/up.png
new file mode 100644
index 0000000..6e9c217
Binary files /dev/null and b/img/up.png differ
diff --git a/save_graph/deep_sarsa_plot.png b/save_graph/deep_sarsa_plot.png
new file mode 100644
index 0000000..7dc00ea
Binary files /dev/null and b/save_graph/deep_sarsa_plot.png differ
diff --git a/save_model/deep_sarsa_pretrained.h5 b/save_model/deep_sarsa_pretrained.h5
new file mode 100644
index 0000000..4a0e512
Binary files /dev/null and b/save_model/deep_sarsa_pretrained.h5 differ