Merge pull request #1861 from ayush-09/game

SpaceInvaders Agent

Author: Yashbhadiyadra

SpaceInvader-Agent/README.md

@@ -0,0 +1,51 @@
+# Reinforcement Learning with SpaceInvaders-v0
+
+This repository contains code for implementing reinforcement learning using the SpaceInvaders-v0 environment from the OpenAI Gym.
+
+## Prerequisites
+
+To run this code, you need the following dependencies:
+
+- Python 3.x
+- Gym: `pip install gym`
+- TensorFlow: `pip install tensorflow`
+- Keras-RL2: `pip install keras-rl2`
+
+## Getting Started
+
+1. Clone the repository: `git clone https://github.com/your_username/your_repository.git`
+2. Navigate to the cloned repository: `cd your_repository`
+
+## Running the Code
+
+1. Open the Python script `space_invaders_rl.py`.
+2. Configure the number of episodes and other parameters as needed.
+3. Run the script: `python space_invaders_rl.py`.
+
+## Understanding the Code
+
+The code performs the following steps:
+
+1. Imports the necessary libraries and initializes the SpaceInvaders-v0 environment.
+2. Runs a specified number of episodes, where each episode represents a game.
+3. Resets the environment for each episode and plays the game until completion.
+4. Renders the environment to visualize the game.
+5. Uses a random policy to select actions.
+6. Accumulates the score and prints the episode number and score.
+7. Closes the environment after all episodes have been completed.
+8. Builds a convolutional neural network model using Keras.
+9. Implements the DQN agent using the Keras-RL2 library.
+10. Compiles the agent with the Adam optimizer.
+11. Trains the agent on the SpaceInvaders-v0 environment.
+12. Tests the trained agent on a few episodes and calculates the average score.
+13. Saves the trained weights of the DQN agent.
+14. Loads the saved weights of the DQN agent.
+
+
+## Acknowledgments
+
+- [OpenAI Gym](https://gym.openai.com/)
+- [Keras-RL2](https://github.com/wau/keras-rl2)
+
+Feel free to modify and adapt this code according to your needs.
+

SpaceInvader-Agent/SpaceInvaders Agent.py

@@ -0,0 +1,70 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Thu Jun  3 13:06:20 2021
+@author: Ayush
+"""
+
+import gym
+env = gym.make("SpaceInvaders-v0")
+
+episodes=10
+
+for episode in range(1,episodes):
+    state = env.reset()
+    done = False
+    score=0
+    
+    while not done:
+        env.render()
+        state, reward,done, info = env.step(env.action_space.sample())
+        score += reward
+    print('Episode: {}\nScore: {}'.format(episode, score))           
+env.close()
+
+# Import Neural Network Packages
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense, Flatten, Conv2D
+from tensorflow.keras.optimizers import Adam
+
+def build_model(height,width,channels,actions):
+    model = Sequential()
+    model.add(Conv2D(32,(8,8),strides=(4,4), activation='relu', input_shape=(3, height,width,channels)))
+    model.add(Conv2D(64,(4,4),strides=(2,2), activation='relu'))
+    model.add(Conv2D(64,(4,4),strides=(2,2), activation='relu'))
+    model.add(Flatten())
+    model.add(Dense(512, activation= 'relu'))
+    model.add(Dense(256,activation='relu'))
+    model.add(Dense(64,activation='relu'))
+    model.add(Dense(actions,activation='linear'))
+    return model
+
+height,width,channels = env.observation_space.shape
+actions = env.action_space.n
+
+model = build_model(height,width,channels,actions)
+
+
+#Importing keras-rl2 reinforcement learning functions
+from rl.agents import DQNAgent
+from rl.memory import SequentialMemory
+from rl.policy import LinearAnnealedPolicy, EpsGreedyQPolicy
+
+def build_agent(model,actions):
+    policy = LinearAnnealedPolicy(EpsGreedyQPolicy(), attr='eps', value_max=1., value_min=.1, value_test=.2, nb_steps=10000)
+    memory = SequentialMemory(limit=2000, window_length=3)
+    dqn = DQNAgent(model=model, memory=memory, policy=policy, enable_dueling_network=True, dueling_type='avg', nb_actions=actions, nb_steps_warmup=1000)
+    return dqn
+
+dqn = build_agent(model, actions)
+
+
+dqn.compile(Adam(lr=0.001))
+
+dqn.fit(env,nb_steps=40000, visualize=True, verbose=1)
+
+import numpy as np
+scores = dqn.test(env,nb_episodes=10,visualize=True)
+print(np.mean(scores.history['episode_reward']))
+
+dqn.save_weights('models/dqn.h5f')
+dqn.load_weights('models/dqn.h5f')