Pendulum with SAC¶
In this notebook we solve the Pendulum environment using SAC. We’ll use a simple multi-layer percentron for our function approximator for the policy and q-function.
This notebook periodically generates GIFs, so that we can inspect how the training is progressing.
After a few hundred episodes, this is what you can expect:
import gymnasium
import jax
import coax
import haiku as hk
import jax.numpy as jnp
from numpy import prod
import optax
# the name of this script
name = 'sac'
# the Pendulum MDP
env = gymnasium.make('Pendulum-v1', render_mode='rgb_array')
env = coax.wrappers.TrainMonitor(env, name=name, tensorboard_dir=f"./data/tensorboard/{name}")
def func_pi(S, is_training):
seq = hk.Sequential((
hk.Linear(8), jax.nn.relu,
hk.Linear(8), jax.nn.relu,
hk.Linear(8), jax.nn.relu,
hk.Linear(prod(env.action_space.shape) * 2, w_init=jnp.zeros),
hk.Reshape((*env.action_space.shape, 2)),
))
x = seq(S)
mu, logvar = x[..., 0], x[..., 1]
return {'mu': mu, 'logvar': logvar}
def func_q(S, A, is_training):
seq = hk.Sequential((
hk.Linear(8), jax.nn.relu,
hk.Linear(8), jax.nn.relu,
hk.Linear(8), jax.nn.relu,
hk.Linear(1, w_init=jnp.zeros), jnp.ravel
))
X = jnp.concatenate((S, A), axis=-1)
return seq(X)
# main function approximators
pi = coax.Policy(func_pi, env)
q1 = coax.Q(func_q, env, action_preprocessor=pi.proba_dist.preprocess_variate)
q2 = coax.Q(func_q, env, action_preprocessor=pi.proba_dist.preprocess_variate)
# target network
q1_targ = q1.copy()
q2_targ = q2.copy()
# experience tracer
tracer = coax.reward_tracing.NStep(n=5, gamma=0.9, record_extra_info=True)
buffer = coax.experience_replay.SimpleReplayBuffer(capacity=25000)
alpha = 0.2
policy_regularizer = coax.regularizers.NStepEntropyRegularizer(pi,
beta=alpha / tracer.n,
gamma=tracer.gamma,
n=[tracer.n])
# updaters (use current pi to update the q-functions and use sampled action in contrast to TD3)
qlearning1 = coax.td_learning.SoftClippedDoubleQLearning(
q1, pi_targ_list=[pi], q_targ_list=[q1_targ, q2_targ],
loss_function=coax.value_losses.mse, optimizer=optax.adam(1e-3),
policy_regularizer=policy_regularizer)
qlearning2 = coax.td_learning.SoftClippedDoubleQLearning(
q2, pi_targ_list=[pi], q_targ_list=[q1_targ, q2_targ],
loss_function=coax.value_losses.mse, optimizer=optax.adam(1e-3),
policy_regularizer=policy_regularizer)
soft_pg = coax.policy_objectives.SoftPG(pi, [q1_targ, q2_targ], optimizer=optax.adam(
1e-3), regularizer=coax.regularizers.NStepEntropyRegularizer(pi,
beta=alpha / tracer.n,
gamma=tracer.gamma,
n=jnp.arange(tracer.n)))
# train
while env.T < 1000000:
s, info = env.reset()
for t in range(env.spec.max_episode_steps):
a = pi(s)
s_next, r, done, truncated, info = env.step(a)
# trace rewards and add transition to replay buffer
tracer.add(s, a, r, done)
while tracer:
buffer.add(tracer.pop())
# learn
if len(buffer) >= 5000:
transition_batch = buffer.sample(batch_size=128)
# init metrics dict
metrics = {}
# flip a coin to decide which of the q-functions to update
qlearning = qlearning1 if jax.random.bernoulli(q1.rng) else qlearning2
metrics.update(qlearning.update(transition_batch))
# delayed policy updates
if env.T >= 7500 and env.T % 4 == 0:
metrics.update(soft_pg.update(transition_batch))
env.record_metrics(metrics)
# sync target networks
q1_targ.soft_update(q1, tau=0.001)
q2_targ.soft_update(q2, tau=0.001)
if done or truncated:
break
s = s_next
# generate an animated GIF to see what's going on
if env.period(name='generate_gif', T_period=10000) and env.T > 5000:
T = env.T - env.T % 10000 # round to 10000s
coax.utils.generate_gif(
env=env, policy=pi, filepath=f"./data/gifs/{name}/T{T:08d}.gif")