Python utils.ReplayBuffer() Examples

The following are 2 code examples of utils.ReplayBuffer(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module utils , or try the search function .
Example #1
Source File: HAC.py    From Hierarchical-Actor-Critic-HAC-PyTorch with MIT License 6 votes vote down vote up 
def __init__(self, k_level, H, state_dim, action_dim, render, threshold, 
                 action_bounds, action_offset, state_bounds, state_offset, lr):
        
        # adding lowest level
        self.HAC = [DDPG(state_dim, action_dim, action_bounds, action_offset, lr, H)]
        self.replay_buffer = [ReplayBuffer()]
        
        # adding remaining levels
        for _ in range(k_level-1):
            self.HAC.append(DDPG(state_dim, state_dim, state_bounds, state_offset, lr, H))
            self.replay_buffer.append(ReplayBuffer())
        
        # set some parameters
        self.k_level = k_level
        self.H = H
        self.action_dim = action_dim
        self.state_dim = state_dim
        self.threshold = threshold
        self.render = render
        
        # logging parameters
        self.goals = [None]*self.k_level
        self.reward = 0
        self.timestep = 0
Example #2
Source File: main.py    From BCQ with MIT License 5 votes vote down vote up 
def train_BCQ(state_dim, action_dim, max_action, device, args):
	# For saving files
	setting = f"{args.env}_{args.seed}"
	buffer_name = f"{args.buffer_name}_{setting}"

	# Initialize policy
	policy = BCQ.BCQ(state_dim, action_dim, max_action, device, args.discount, args.tau, args.lmbda, args.phi)

	# Load buffer
	replay_buffer = utils.ReplayBuffer(state_dim, action_dim, device)
	replay_buffer.load(f"./buffers/{buffer_name}")
	
	evaluations = []
	episode_num = 0
	done = True 
	training_iters = 0
	
	while training_iters < args.max_timesteps: 
		pol_vals = policy.train(replay_buffer, iterations=int(args.eval_freq), batch_size=args.batch_size)

		evaluations.append(eval_policy(policy, args.env, args.seed))
		np.save(f"./results/BCQ_{setting}", evaluations)

		training_iters += args.eval_freq
		print(f"Training iterations: {training_iters}")


# Runs policy for X episodes and returns average reward
# A fixed seed is used for the eval environment