def train(env_name, seed=42, timesteps=1, epsilon_decay_last_step=1000,
er_capacity=1e4, batch_size=16, lr=1e-3, gamma=1.0, update_target=16,
exp_name='test', init_timesteps=100, save_every_steps=1e4, arch='nature',
dueling=False, play_steps=2, n_jobs=2):
"""
Main training function. Calls the subprocesses to get experience and
train the network.
"""
# Multiprocessing method
mp.set_start_method('spawn')
# Get PyTorch device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Set random seed for PyTorch
torch.manual_seed(seed)
torch.cuda.manual_seed(seed)
# Create logger
logger = Logger(exp_name, loggers=['tensorboard'])
# Create the Q network
_env = make_env(env_name, seed)
net = QNetwork(_env.observation_space, _env.action_space, arch=arch, dueling=dueling).to(device)
# Create the target network as a copy of the Q network
target_net = copy.deepcopy(net)
# Create buffer and optimizer
buffer = ExperienceReplay(capacity=int(er_capacity))
optimizer = optim.Adam(net.parameters(), lr=lr)
scheduler = StepLR(optimizer, step_size=LR_STEPS, gamma=0.99)
# Multiprocessing queue
obs_queue = mp.Queue(maxsize=n_jobs)
transition_queue = mp.Queue(maxsize=n_jobs)
workers, action_queues = [], []
for i in range(n_jobs):
action_queue = mp.Queue(maxsize=1)
_seed = seed + i * 1000
play_proc = mp.Process(target=play_func, args=(i, env_name, obs_queue, transition_queue, action_queue, _seed))
play_proc.start()
workers.append(play_proc)
action_queues.append(action_queue)
# Vars to keep track of performances and time
timestep = 0
current_reward, current_len = np.zeros(play_steps), np.zeros(play_steps, dtype=np.int64)
current_time = [time.time() for _ in range(play_steps)]
# Training loop
while timestep < timesteps:
# Compute the current epsilon
epsilon = EPSILON_STOP + max(0, (EPSILON_START - EPSILON_STOP)*(epsilon_decay_last_step-timestep)/epsilon_decay_last_step)
logger.log_kv('internals/epsilon', epsilon, timestep)
# Gather observation N_STEPS
ids, obs_batch = zip(*[obs_queue.get() for _ in range(play_steps)])
# Pre-process observation_batch for PyTorch
obs_batch = torch.from_numpy(np.array(obs_batch)).to(device)
# Select greedy action from policy, apply epsilon-greedy selection
greedy_actions = net(obs_batch).argmax(dim=1).cpu().detach().numpy()
probs = torch.rand(greedy_actions.shape)
actions = np.where(probs < epsilon, _env.action_space.sample(), greedy_actions)
# Send actions
for id, action in zip(ids, actions):
action_queues[id].put(action)
# Add transitions to experience replay
transitions = [transition_queue.get() for _ in range(play_steps)]
buffer.pushTransitions(transitions)
# Check if we need to update rewards, time and lengths
_, _, _, reward, done, _ = zip(*transitions)
current_reward += reward
current_len += 1
for i, done in enumerate(done):
if done:
# Log quantities
logger.log_kv('performance/return', current_reward[i], timestep)
logger.log_kv('performance/length', current_len[i], timestep)
logger.log_kv('performance/speed', current_len[i] / (time.time() - current_time[i]), timestep)
# Reset counters
current_reward[i] = 0.0
current_len[i] = 0
current_time[i] = time.time()
# Update number of steps
timestep += play_steps
# Check if we are in the warm-up phase, otherwise go on with policy update
if timestep < init_timesteps:
continue
# Learning rate upddate and log
scheduler.step()
logger.log_kv('internals/lr', scheduler.get_lr()[0], timestep)
# Clear grads
optimizer.zero_grad()
# Get a batch from experience replay
batch = buffer.sampleTransitions(batch_size)
def batch_preprocess(batch_item):
return torch.tensor(batch_item, dtype=(torch.long if isinstance(batch_item[0], np.int64) else None)).to(device)
ids, states_batch, actions_batch, rewards_batch, done_batch, next_states_batch = map(batch_preprocess, zip(*batch))
# Compute the loss function
state_action_values = net(states_batch).gather(1, actions_batch.unsqueeze(-1)).squeeze(-1)
next_state_values = target_net(next_states_batch).max(1)[0]
next_state_values[done_batch] = 0.0
expected_state_action_values = next_state_values.detach() * gamma + rewards_batch
loss = F.mse_loss(state_action_values, expected_state_action_values)
logger.log_kv('internals/loss', loss.item(), timestep)
loss.backward()
# Clip the gradients to avoid to abrupt changes (this is equivalent to Huber Loss)
for param in net.parameters():
param.grad.data.clamp_(-1, 1)
optimizer.step()
if timestep % update_target == 0:
target_net.load_state_dict(net.state_dict())
# Check if we need to save a checkpoint
if timestep % save_every_steps == 0:
torch.save(net.get_extended_state(), exp_name + '.pth')
# Ending
for i, worker in enumerate(workers):
action_queues[i].put(None)
worker.join()
def train(env_name,
arch,
timesteps=1,
init_timesteps=0,
seed=42,
er_capacity=1,
epsilon_start=1.0,
epsilon_stop=0.05,
epsilon_decay_stop=1,
batch_size=16,
target_sync=16,
lr=1e-3,
gamma=1.0,
dueling=False,
play_steps=1,
lr_steps=1e4,
lr_gamma=0.99,
save_steps=5e4,
logger=None,
experiment_name='test'):
"""
Main training function. Calls the subprocesses to get experience and
train the network.
"""
# Casting params which are expressable in scientific notation
def int_scientific(x):
return int(float(x))
timesteps, init_timesteps = map(int_scientific,
[timesteps, init_timesteps])
lr_steps, epsilon_decay_stop = map(int_scientific,
[lr_steps, epsilon_decay_stop])
er_capacity, target_sync, save_steps = map(
int_scientific, [er_capacity, target_sync, save_steps])
lr = float(lr)
# Multiprocessing method
mp.set_start_method('spawn')
# Get PyTorch device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Create the Q network
_env = make_env(env_name, seed)
net = QNetwork(_env.observation_space,
_env.action_space,
arch=arch,
dueling=dueling).to(device)
# Create the target network as a copy of the Q network
tgt_net = ptan.agent.TargetNet(net)
# Create buffer and optimizer
buffer = ptan.experience.ExperienceReplayBuffer(experience_source=None,
buffer_size=er_capacity)
optimizer = optim.Adam(net.parameters(), lr=lr)
scheduler = StepLR(optimizer, step_size=lr_steps, gamma=0.99)
# Multiprocessing queue
epsilon_schedule = (epsilon_start, epsilon_stop, epsilon_decay_stop)
exp_queue = mp.Queue(maxsize=play_steps * 2)
play_proc = mp.Process(target=play_func,
args=(env_name, net, exp_queue, seed, timesteps,
epsilon_schedule, gamma))
play_proc.start()
# Main training loop
timestep = 0
while play_proc.is_alive() and timestep < timesteps:
timestep += play_steps
# Query the environments and log results if the episode has ended
for _ in range(play_steps):
exp, info = exp_queue.get()
if exp is None:
play_proc.join()
break
buffer._add(exp)
logger.log_kv('internals/epsilon', info['epsilon'][0],
info['epsilon'][1])
if 'ep_reward' in info.keys():
logger.log_kv('performance/return', info['ep_reward'],
timestep)
logger.log_kv('performance/length', info['ep_length'],
timestep)
logger.log_kv('performance/speed', info['speed'], timestep)
# Check if we are in the starting phase
if len(buffer) < init_timesteps:
continue
scheduler.step()
logger.log_kv('internals/lr', scheduler.get_lr()[0], timestep)
# Get a batch from experience replay
optimizer.zero_grad()
batch = buffer.sample(batch_size * play_steps)
# Unpack the batch
states, actions, rewards, dones, next_states = unpack_batch(batch)
states_v = torch.tensor(states).to(device)
next_states_v = torch.tensor(next_states).to(device)
actions_v = torch.tensor(actions).to(device)
rewards_v = torch.tensor(rewards).to(device)
done_mask = torch.ByteTensor(dones).to(device)
# Optimize defining the loss function
state_action_values = net(states_v).gather(
1, actions_v.unsqueeze(-1)).squeeze(-1)
next_state_values = tgt_net.target_model(next_states_v).max(1)[0]
next_state_values[done_mask] = 0.0
expected_state_action_values = next_state_values.detach(
) * gamma + rewards_v
loss = F.mse_loss(state_action_values, expected_state_action_values)
logger.log_kv('internals/loss', loss.item(), timestep)
loss.backward()
# Clip the gradients to avoid to abrupt changes (this is equivalent to Huber Loss)
for param in net.parameters():
param.grad.data.clamp_(-1, 1)
optimizer.step()
# Check if the target network need to be synched
if timestep % target_sync == 0:
tgt_net.sync()
# Check if we need to save a checkpoint
if timestep % save_steps == 0:
torch.save(net.get_extended_state(), experiment_name + '.pth')
