Q-target fissi

Stai per addestrare il tuo Lunar Lander usando Q-target fissi. Come prerequisito, devi istanziare sia la rete online (che sceglie l’azione) sia la rete target (usata per il calcolo del TD-target).

Devi anche implementare una funzione update_target_network che potrai usare a ogni step di training. La rete target non viene aggiornata con la discesa del gradiente; invece, update_target_network spinge i suoi pesi verso quelli della Q-network di una piccola quantità, garantendo che rimanga abbastanza stabile nel tempo.

Nota che, solo per questo esercizio, usi una rete molto piccola così possiamo stampare e ispezionare facilmente il suo state dictionary. Ha un solo livello nascosto di dimensione due; anche lo spazio delle azioni e lo spazio degli stati hanno dimensione 2.

La funzione print_state_dict() è disponibile nel tuo ambiente per stampare lo state dict.

Questo esercizio fa parte del corso

Deep Reinforcement Learning in Python

Visualizza il corso

Istruzioni dell'esercizio

Ottieni lo .state_dict() sia per la rete target sia per la rete online.
Aggiorna lo state dict della rete target facendo la media pesata tra i parametri della rete online e quelli della rete target, usando tau come peso per la rete online.
Carica lo state dict aggiornato nella rete target.

Esercizio pratico interattivo

Prova a risolvere questo esercizio completando il codice di esempio.

def update_target_network(target_network, online_network, tau):
    # Obtain the state dicts for both networks
    target_net_state_dict = ____
    online_net_state_dict = ____
    for key in online_net_state_dict:
        # Calculate the updated state dict for the target network
        target_net_state_dict[key] = (online_net_state_dict[____] * ____ + target_net_state_dict[____] * ____)
        # Load the updated state dict into the target network
        target_network.____
    return None
  
print("online network weights:", print_state_dict(online_network))
print("target network weights (pre-update):", print_state_dict(target_network))
update_target_network(target_network, online_network, .001)
print("target network weights (post-update):", print_state_dict(target_network))

Modifica ed esegui il codice

Questo esercizio fa parte del corso

Deep Reinforcement Learning in Python

AvançadoNível de habilidade

4.8+

Inizia il corso gratis

Discover how deep reinforcement learning improves upon traditional Reinforcement Learning while studying and implementing your first Deep Q Learning algorithm.

Exercise 1: Introduction to deep reinforcement learning Exercise 2: Environment and neural network setup Exercise 3: DRL training loop Exercise 4: Introduction to deep Q learning Exercise 5: Deep learning and DQN Exercise 6: The Q-Network architecture Exercise 7: Instantiating the Q-Network Exercise 8: The barebone DQN algorithm Exercise 9: Barebone DQN action selection Exercise 10: Barebone DQN loss function Exercise 11: Training the barebone DQN

Dive into Deep Q-learning by implementing the original DQN algorithm, featuring Experience Replay, epsilon-greediness and fixed Q-targets. Beyond DQN, you will then explore two fascinating extensions that improve the performance and stability of Deep Q-learning: Double DQN and Prioritized Experience Replay.

Exercise 1: DQN con experience replay Exercise 2: La coda a due estremità Exercise 3: Experience replay buffer Exercise 4: DQN con experience replay Exercise 5: L'algoritmo DQN completo Exercise 6: Epsilon-greediness Exercise 7: Q-target fissi

Esercizio in corso

Exercise 8: Implementare l'algoritmo DQN completo Exercise 9: Double DQN Exercise 10: Rete online e rete target in DDQN Exercise 11: Addestrare il Double DQN Exercise 12: Prioritized experience replay Exercise 13: Prioritized experience replay buffer Exercise 14: Campionare dal buffer PER Exercise 15: DQN con prioritized experience replay

Learn about the foundational concepts of policy gradient methods found in DRL. You will begin with the policy gradient theorem, which forms the basis for these methods. Then, you will implement the REINFORCE algorithm, a powerful approach to learning policies. The chapter will then guide you through Actor-Critic methods, focusing on the Advantage Actor-Critic (A2C) algorithm, which combines the strengths of both policy gradient and value-based methods to enhance learning efficiency and stability.

Exercise 1: Introduction to policy gradient Exercise 2: The policy network architecture Exercise 3: Working with discrete distributions Exercise 4: Policy gradient and REINFORCE Exercise 5: Action selection in REINFORCE Exercise 6: Training the REINFORCE algorithm Exercise 7: Advantage Actor Critic Exercise 8: Critic network Exercise 9: Actor Critic loss calculations Exercise 10: Training the A2C algorithm

Explore Proximal Policy Optimization (PPO) for robust DRL performance. Next, you will examine using an entropy bonus in PPO, which encourages exploration by preventing premature convergence to deterministic policies. You'll also learn about batch updates in policy gradient methods. Finally, you will learn about hyperparameter optimization with Optuna, a powerful tool for optimizing performance in your DRL models.

Exercise 1: Proximal policy optimization Exercise 2: The clipped probability ratio Exercise 3: The clipped surrogate objective function Exercise 4: Entropy bonus and PPO Exercise 5: Entropy playground Exercise 6: Training the PPO algorithm Exercise 7: Batch updates in policy gradient Exercise 8: Minibatch and DRL Exercise 9: A2C with batch updates Exercise 10: Hyperparameter optimization with Optuna Exercise 11: Hyperparameter or not?Exercise 12: Hands-on with Optuna Exercise 13: Congratulations!