tdpeuter/2026SEL3-project-Brittle_Star_Controller_Modularity
tdpeuter/2026SEL3-project-Brittle_Star_Controller_Modularity
Archived
1
Fork 0
Code Issues Releases Activity

fix: uniform naming

Browse source
This commit is contained in:
cedric 2026-04-03 08:11:50 +00:00
parent 1c4e40b5fb
commit 3104678f96
1 changed files with 6 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

12
src/ppo.py
View file

@ -85,9 +85,9 @@ that are now not in the same scope
@partial(jax.jit, static_argnums=(0, 1, 2, 3, 4)) @partial(jax.jit, static_argnums=(0, 1, 2, 3, 4))
def get_action_and_value2( def get_action_and_value(
input_apply, sensor_apply,
action_apply, actor_apply,
message_passer, message_passer,
critic_apply, critic_apply,
feature_extractor_apply, feature_extractor_apply,
@ -95,10 +95,10 @@ def get_action_and_value2(
x: jnp.ndarray, x: jnp.ndarray,
action: jnp.ndarray, action: jnp.ndarray,
): ):
hidden_sensor = input_apply(params["sensor_params"], x) hidden_sensor = sensor_apply(params["sensor_params"], x)
hidden_critic = feature_extractor_apply(params["feature_extractor_params"], x) hidden_critic = feature_extractor_apply(params["feature_extractor_params"], x)
hidden_sensor = message_passer(hidden_sensor) hidden_sensor = message_passer(hidden_sensor)
mean, log_std = action_apply(params["actor_params"], hidden_sensor) mean, log_std = actor_apply(params["actor_params"], hidden_sensor)
std = jnp.exp(log_std) std = jnp.exp(log_std)
logprob = -0.5 * (((action - mean) / std) ** 2 + 2 * log_std + jnp.log(2 * jnp.pi)).sum(-1) logprob = -0.5 * (((action - mean) / std) ** 2 + 2 * log_std + jnp.log(2 * jnp.pi)).sum(-1)
@ -122,7 +122,7 @@ def ppo_loss(
critic_apply, critic_apply,
feature_extractor_apply, feature_extractor_apply,
): ):
newlogprob, entropy, newvalue = get_action_and_value2( newlogprob, entropy, newvalue = get_action_and_value(
sensor_apply, sensor_apply,
actor_apply, actor_apply,
message_passer, message_passer,