---
tags:
- LunarLanderContinuous-v2
- reinforce
- reinforcement-learning
- custom-implementation
model-index:
- name: REINFORCE-LunarLanderContinuous-v2
  results:
  - task:
      type: reinforcement-learning
      name: reinforcement-learning
    dataset:
      name: LunarLanderContinuous-v2
      type: LunarLanderContinuous-v2
    metrics:
    - type: mean_reward
      value: 264.10 +/- 37.17
      name: mean_reward
      verified: false
---

# **Reinforce** Agent playing **LunarLanderContinuous-v2**
This is a custom agent. Performance has been measured over 900 episodes.
To try the agent, user needs to import the ParameterisedPolicy class from the Agent_class.py file. </br>
Training progress:
![training](training_graph.jpg)

Numbers on X axis are average over 40 episodes, each lasting for about 500 timesteps on average. So in total the agent was trained over about 5e6 timesteps.
Learning rate decay schedule: <code>torch.optim.lr_scheduler.StepLR(opt, step_size=4000, gamma=0.7)</code>

Minimal code to use the agent:</br>
<pre><code>
import gym</br>
</br>
env_name = 'LunarLanderContinuous-v2'</br>
env = gym.make(env_name)</br>
agent = torch.load('best_models/best_reinforce_lunar_lander_cont_model_269.402.pt')</br>
render = True</br>
observation = env.reset()</br>
while True:</br>
    if render:</br>
        env.render()</br>
    action = agent.act(observation)</br>
    observation, reward, done, info = env.step(action)</br>
    </br>
    if done:</br>
        break</br>
env.close()</br>
</code></pre>