def main(args): hidden_units = args.hidden_units msg_dim = args.msg_dim model_path = os.getcwd() + "/" + args.model_dir ray.init(log_to_driver=False) env_test_instance = gym.make('BipedalWalker-v3') if args.baseline: from smp.baseline import TD3Net action_dimension = copy(env_test_instance.action_space.shape[0]) else: from smp.smp import TD3Net action_dimension = 1 model_kwargs = { # action dimension for modular actions 'action_dimension': action_dimension, 'min_action': copy(env_test_instance.action_space.low)[0], 'max_action': copy(env_test_instance.action_space.high)[0], 'msg_dimension': msg_dim, 'fix_sigma': True, 'hidden_units': hidden_units } del env_test_instance manager = SampleManager(TD3Net, 'BipedalWalker-v3', num_parallel=(os.cpu_count() - 1), total_steps=150, action_sampling_type="continuous_normal_diagonal", is_tf=True, model_kwargs=model_kwargs) manager.load_model(model_path) manager.test(200, test_episodes=5, render=True) ray.shutdown()
"model": QNet, "environment": "CartPole-v1", "num_parallel": 1, "total_steps": 1000, "model_kwargs": model_kwargs, } # Initialize ray.init(log_to_driver=False) manager = SampleManager(**kwargs) # Where to load your results from loading_path = os.getcwd() + "/progress_CartPole" # Load model manager.load_model(loading_path) print("done") print("testing optimized agent") manager.test( 1000, test_episodes=10, render=True, do_print=True, evaluation_measure="time_and_reward", ) print('Prepare LunarLander') env = gym.make("LunarLander-v2") model_kwargs = {"layers": [32, 32, 32], "num_actions": env.action_space.n}
def train_td3(args, model, action_dimension=None): print(args) tf.keras.backend.set_floatx('float32') ray.init(log_to_driver=False) # hyper parameters buffer_size = args.buffer_size # 10e6 in their repo, not possible with our ram epochs = args.epochs saving_path = os.getcwd() + "/" + args.saving_dir saving_after = 5 sample_size = args.sample_size optim_batch_size = args.batch_size gamma = args.gamma test_steps = 100 # 1000 in their repo policy_delay = 2 rho = .046 policy_noise = args.policy_noise policy_noise_clip = .5 msg_dim = args.msg_dim # 32 in their repo learning_rate = args.learning_rate save_args(args, saving_path) env_test_instance = gym.make('BipedalWalker-v3') if action_dimension is None: action_dimension = copy(env_test_instance.action_space.shape[0]) model_kwargs = { # action dimension for modular actions 'action_dimension': action_dimension, 'min_action': copy(env_test_instance.action_space.low)[0], 'max_action': copy(env_test_instance.action_space.high)[0], 'msg_dimension': msg_dim, 'fix_sigma': True, 'hidden_units': args.hidden_units } del env_test_instance manager = SampleManager(model, 'BipedalWalker-v3', num_parallel=(os.cpu_count() - 1), total_steps=150, action_sampling_type="continuous_normal_diagonal", is_tf=True, model_kwargs=model_kwargs) optim_keys = [ 'state', 'action', 'reward', 'state_new', 'not_done', ] manager.initialize_buffer(buffer_size, optim_keys) manager.initialize_aggregator(path=saving_path, saving_after=saving_after, aggregator_keys=["loss", "reward"]) agent = manager.get_agent() optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate) # fill buffer print("Filling buffer before training..") while len(manager.buffer.buffer[ manager.buffer.keys[0]]) < manager.buffer.size: # Gives you state action reward trajectories data = manager.get_data() manager.store_in_buffer(data) # track time while training timer = time.time() last_t = timer target_agent = manager.get_agent() for e in range(epochs): # off policy sample_dict = manager.sample(sample_size, from_buffer=True) print(f"collected data for: {sample_dict.keys()}") # cast values to float32 and create data dict sample_dict['state'] = tf.cast(sample_dict['state'], tf.float32) sample_dict['action'] = tf.cast(sample_dict['action'], tf.float32) sample_dict['reward'] = tf.cast(sample_dict['reward'], tf.float32) sample_dict['state_new'] = tf.cast(sample_dict['state_new'], tf.float32) sample_dict['not_done'] = tf.cast(sample_dict['not_done'], tf.float32) data_dict = dict_to_dict_of_datasets(sample_dict, batch_size=optim_batch_size) total_loss = 0 for state, action, reward, state_new, not_done in \ zip(data_dict['state'], data_dict['action'], data_dict['reward'], data_dict['state_new'], data_dict['not_done']): action_new = target_agent.act(state_new) # add noise to action_new action_new = action_new + tf.clip_by_value( tf.random.normal(action_new.shape, 0., policy_noise), -policy_noise_clip, policy_noise_clip) # clip action_new to action space action_new = tf.clip_by_value( action_new, manager.env_instance.action_space.low, manager.env_instance.action_space.high) # calculate target with double-Q-learning state_action_new = tf.concat([state_new, action_new], axis=-1) q_values0 = target_agent.model.critic0(state_action_new) q_values1 = target_agent.model.critic1(state_action_new) q_values = tf.concat([q_values0, q_values1], axis=-1) q_targets = tf.squeeze(tf.reduce_min(q_values, axis=-1)) critic_target = reward + gamma * not_done * q_targets state_action = tf.concat([state, action], axis=-1) # update critic 0 with tf.GradientTape() as tape: q_output = agent.model.critic0(state_action) loss = tf.keras.losses.MSE(tf.squeeze(critic_target), tf.squeeze(q_output)) total_loss += loss gradients = tape.gradient(loss, agent.model.critic0.trainable_variables) optimizer.apply_gradients( zip(gradients, agent.model.critic0.trainable_variables)) # update critic 1 with tf.GradientTape() as tape: q_output = agent.model.critic1(state_action) loss = tf.keras.losses.MSE(tf.squeeze(critic_target), tf.squeeze(q_output)) total_loss += loss gradients = tape.gradient(loss, agent.model.critic1.trainable_variables) optimizer.apply_gradients( zip(gradients, agent.model.critic1.trainable_variables)) # update actor with delayed policy update if e % policy_delay == 0: with tf.GradientTape() as tape: actor_output = agent.model.actor(state) action = reparam_action(actor_output, agent.model.action_dimension, agent.model.min_action, agent.model.max_action) state_action = tf.concat([state, action], axis=-1) q_val = agent.model.critic0(state_action) actor_loss = -tf.reduce_mean(q_val) total_loss += actor_loss actor_gradients = tape.gradient( actor_loss, agent.model.actor.trainable_variables) optimizer.apply_gradients( zip(actor_gradients, agent.model.actor.trainable_variables)) # Update agent manager.set_agent(agent.get_weights()) agent = manager.get_agent() if e % policy_delay == 0: # Polyak averaging new_weights = list(rho * np.array(target_agent.get_weights()) + (1. - rho) * np.array(agent.get_weights())) target_agent.set_weights(new_weights) reward = manager.test(test_steps, evaluation_measure="reward") manager.update_aggregator(loss=total_loss, reward=reward) print( f"epoch ::: {e} loss ::: {total_loss} avg reward ::: {np.mean(reward)}" ) if e % saving_after == 0: manager.save_model(saving_path, e) # needed time and remaining time estimation current_t = time.time() time_needed = (current_t - last_t) / 60. time_remaining = (current_t - timer) / 60. / (e + 1) * (epochs - (e + 1)) print( 'Finished epoch %d of %d. Needed %1.f min for this epoch. Estimated time remaining: %.1f min' % (e + 1, epochs, time_needed, time_remaining)) last_t = current_t manager.load_model(saving_path) print("done") print("testing optimized agent") manager.test(test_steps, test_episodes=10, render=True) ray.shutdown()
# positive critic loss for gradient descent with MSE critic_loss = tf.reduce_mean((mc - agent.v(state))**2) critic_gradients = tape.gradient( critic_loss, agent.model.critic.trainable_variables) optimizer.apply_gradients( zip(critic_gradients, agent.model.critic.trainable_variables)) total_loss += actor_loss + critic_loss # Update the agent manager.set_agent(agent.get_weights()) agent = manager.get_agent() reward = manager.test(test_steps, evaluation_measure="reward") manager.update_aggregator(loss=total_loss, reward=reward) # print progress print( f"epoch ::: {e} loss ::: {total_loss} avg env steps ::: {np.mean(reward)}" ) if e % saving_after == 0: # you can save models manager.save_model(saving_path, e) # and load mmodels manager.load_model(saving_path) print("done") print("testing optimized agent") manager.test(test_steps, test_episodes=10, render=True)
# ---------------------- IO ---------------------- """ This section saves plots of the training process, writes some details to csv, and allow to continue training from an existing models. """ saving_path = os.getcwd() + "/" + env_name manager.initialize_aggregator(path=saving_path, saving_after=5, aggregator_keys=["loss", 'reward', 'time']) results_file_name = env_name + f"/results_{'rnd' if use_rnd else 'base'}.csv" with open(results_file_name, 'a') as fd: fd.write('epoch,loss,reward,rnd_loss,steps\n') if continue_from_saved_model: agent, epoch_offset = manager.load_model(saving_path) else: agent = manager.get_agent() epoch_offset = 0 # ====================== Training ====================== rewards = [] print('TRAINING') for e in range(epoch_offset, max_episodes + epoch_offset): e += 1 t = time.time() # ====================== Setup ====================== """ In the Setup phase we will: 1. sample trajectories