From b4fa5624143643a3b7d641e4fd2c91f4af8cad74 Mon Sep 17 00:00:00 2001
From: Anish Muthali <anishmuthali@gmail.com>
Date: Mon, 21 Oct 2019 20:50:27 -0700
Subject: [PATCH 1/3] Added basic test model

---
 DQN_model.py | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
 create mode 100644 DQN_model.py

diff --git a/DQN_model.py b/DQN_model.py
new file mode 100644
index 0000000..0682900
--- /dev/null
+++ b/DQN_model.py
@@ -0,0 +1,31 @@
+import os
+import gym
+import gym_donkeycar
+import numpy as np
+from stable_baselines.common.vec_env import DummyVecEnv
+from stable_baselines.deepq.policies import MlpPolicy
+from stable_baselines import DQN
+
+#SET UP ENVIRONMENT
+os.environ['DONKEY_SIM_PATH'] = f"./DonkeySimMac/donkey_sim.app/Contents/MacOS/donkey_sim"
+os.environ['DONKEY_SIM_PORT'] = str(9091)
+os.environ['DONKEY_SIM_HEADLESS'] = str(1) # "1" is headless
+
+env = gym.make("donkey-warehouse-v0")
+#gym.make("donkey-generated-roads-v0")
+
+timesteps = 100000 # Set this to a reasonable number
+model_name = "dqn_model" # Change the model name to your preferences
+training = True # Change this to test or use the model
+
+if training:
+    model = DQN(MlpPolicy, env, verbose=1)
+    model.learn(total_timesteps=timesteps)
+    model.save(model_name)
+else:
+    model = DQN.load(model_name)
+    obv = env.reset()
+    for t in range(10000):
+        action, _states = model.predict(obv) # drive straight with small speed
+    # execute the action
+        obv, reward, done, info = env.step(action)

From 106e4fc05d2e9fbb208ee4951cca03cdb7ead0e0 Mon Sep 17 00:00:00 2001
From: Ethan Burrell <30448729+ethanburrell@users.noreply.github.com>
Date: Mon, 18 Nov 2019 19:46:36 -0800
Subject: [PATCH 2/3] Update README.md

---
 README.md | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/README.md b/README.md
index df8801e..ab2928c 100644
--- a/README.md
+++ b/README.md
@@ -55,6 +55,7 @@ https://github.com/tawnkramer/gym-donkeycar/tree/master/examples
 Download program for your computer:
 https://github.com/tawnkramer/gym-donkeycar/releases
 
+```
 import os
 import gym
 import gym_donkeycar
@@ -73,3 +74,5 @@ for t in range(100):
     action = np.array([0.0,0.5]) # drive straight with small speed
 # execute the action
 obv, reward, done, info = env.step(action)
+
+```

From 70576821c277f63405c49cd28224b29d0d961a02 Mon Sep 17 00:00:00 2001
From: Anish Muthali <anishmuthali@gmail.com>
Date: Thu, 21 Nov 2019 20:43:42 -0800
Subject: [PATCH 3/3] Added gitignore + bad DQN model

---
 .gitignore   |  2 ++
 DQN_model.py | 15 ++++++++-------
 2 files changed, 10 insertions(+), 7 deletions(-)
 create mode 100644 .gitignore

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..bfc8844
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,2 @@
+unitylog.txt
+env
diff --git a/DQN_model.py b/DQN_model.py
index 0682900..1365cef 100644
--- a/DQN_model.py
+++ b/DQN_model.py
@@ -3,27 +3,28 @@
 import gym_donkeycar
 import numpy as np
 from stable_baselines.common.vec_env import DummyVecEnv
-from stable_baselines.deepq.policies import MlpPolicy
-from stable_baselines import DQN
+from stable_baselines.ddpg.policies import CnnPolicy
+from stable_baselines import DDPG
 
 #SET UP ENVIRONMENT
-os.environ['DONKEY_SIM_PATH'] = f"./DonkeySimMac/donkey_sim.app/Contents/MacOS/donkey_sim"
+os.environ['DONKEY_SIM_PATH'] = f"../DonkeySimMac/donkey_sim.app/Contents/MacOS/donkey_sim"
 os.environ['DONKEY_SIM_PORT'] = str(9091)
-os.environ['DONKEY_SIM_HEADLESS'] = str(1) # "1" is headless
+os.environ['DONKEY_SIM_HEADLESS'] = str(0) # "1" is headless
 
 env = gym.make("donkey-warehouse-v0")
 #gym.make("donkey-generated-roads-v0")
 
 timesteps = 100000 # Set this to a reasonable number
-model_name = "dqn_model" # Change the model name to your preferences
+model_name = "ddpg_model" # Change the model name to your preferences
 training = True # Change this to test or use the model
 
 if training:
-    model = DQN(MlpPolicy, env, verbose=1)
+    vec_env = DummyVecEnv([lambda: env])
+    model = DDPG(CnnPolicy, vec_env, verbose=1)
     model.learn(total_timesteps=timesteps)
     model.save(model_name)
 else:
-    model = DQN.load(model_name)
+    model = DDPG.load(model_name)
     obv = env.reset()
     for t in range(10000):
         action, _states = model.predict(obv) # drive straight with small speed