fix: custom reward function dependent on env reward + extensions

configs/environment/directed_locomotion.yaml

@@ -2,11 +2,11 @@
 # Baseline task setting.
 
 task: DIRECTED_LOCOMOTION
-simulation_time: 5.0
+simulation_time: 5000.0
 num_physics_steps_per_control_step: 10
 time_scale: 2
 camera_ids: [0, 1]
 render_size: [480, 640]
 joint_randomization_noise_scale: 0.0
-target_distance: 3.0
+target_distance: 0.6
 light_perlin_noise_scale: 0

configs/logging/wandb_enabled.yaml

@@ -2,7 +2,7 @@
 # For production/cloud experiments with weights synced.
 
 track: true
-wandb_project_name: "PPO-Modularity"
+wandb_project_name: "PPO-Modularity - reward engineering"
 wandb_entity: "SEL3-2026-Groep-4"
 capture_video: false
 save_model: true

configs/morphology/2_arms.yaml

@@ -0,0 +1,5 @@
+# 2 Arms Morphology Configuration
+
+segments_per_arm: [4, 4]
+use_p_control: true
+use_torque_control: false

configs/ppo/debug.yaml

@@ -1,5 +1,5 @@
 learning_rate: 0.0003
-total_timesteps: 409600 
+total_timesteps: 409600
 num_envs: 32
 num_steps: 32
 anneal_lr: true

configs/ppo/fast.yaml

@@ -3,7 +3,7 @@
 
 learning_rate: 0.0005
 total_timesteps: 500000
-num_envs: 8
+num_envs: 32
 num_steps: 128
 anneal_lr: true
 gamma: 0.99

configs/ppo/fast2.yaml

@@ -0,0 +1,16 @@
+learning_rate: 0.0003
+total_timesteps: 1228800
+num_envs: 32
+num_steps: 64
+anneal_lr: true
+gamma: 0.99
+gae_lambda: 0.95
+num_minibatches: 32
+update_epochs: 4
+norm_adv: true
+clip_coef: 0.2
+clip_vloss: true
+ent_coef: 0.005
+vf_coef: 1.0
+max_grad_norm: 0.5
+target_kl: null