roll back to 1930 and roll back rewards and start polishing. Add some…

… more random rotation to car.
Kaiyotech · Dec 30, 2022 · 3481387 · 3481387
1 parent 3eb8e4e
commit 3481387
Show file tree

Hide file tree

Showing 4 changed files with 13 additions and 17 deletions.
diff --git a/Constants_recovery.py b/Constants_recovery.py
@@ -2,5 +2,5 @@
 TIME_HORIZON = 2 # horizon in seconds
 T_STEP = FRAME_SKIP / 120   # real time per rollout step
 ZERO_SUM = False
-STEP_SIZE = 500_000
+STEP_SIZE = 1_000_000
 DB_NUM = 8
diff --git a/learner_recovery.py b/learner_recovery.py
@@ -43,22 +43,22 @@
     fps = 120 / frame_skip
     gamma = np.exp(np.log(0.5) / (fps * half_life_seconds))
     config = dict(
-        actor_lr=1e-4,
-        critic_lr=1e-4,
+        actor_lr=1e-5,
+        critic_lr=1e-5,
         n_steps=Constants_recovery.STEP_SIZE,
         batch_size=250_000,
         minibatch_size=125_000,
         epochs=30,
         gamma=gamma,
-        save_every=10,
+        save_every=5,
         model_every=1000,
-        ent_coef=0.015,
+        ent_coef=0.01,
     )
 
-    run_id = "recovery_run6.04"
+    run_id = "recovery_run6.05"
     wandb.login(key=os.environ["WANDB_KEY"])
     logger = wandb.init(dir="./wandb_store",
-                        name="Recovery_Run6.04",
+                        name="Recovery_Run6.05",
                         project="Opti",
                         entity="kaiyotech",
                         id=run_id,
@@ -84,13 +84,10 @@
                                                               velocity_pb_w=0.01,
                                                               boost_gain_w=0.25,
                                                               punish_boost=True,
-                                                              touch_ball_w=2,
+                                                              touch_ball_w=3,
                                                               boost_remain_touch_w=.75,
                                                               touch_grass_w=-0.03,
                                                               supersonic_bonus_vpb_w=0.02,
-                                                              zero_touch_grass_if_ss=True,
-                                                              turtle_w=-0.04,
-                                                              final_reward_ball_dist_w=1,
                                                               ),
                                         lambda: CoyoteAction(),
                                         save_every=logger.config.save_every * 3,

diff --git a/setter.py b/setter.py
@@ -174,13 +174,15 @@ def __init__(self, mode, end_object_choice=None):
                             AugmentSetter(ReplaySetter(low_recovery_replays[i], random_boost=False,
                                                        zero_ball_weight=0.8,
                                                        zero_car_weight=0.2,
-                                                       rotate_car_weight=0.3,
+                                                       rotate_car_weight=0.2,
+                                                       backward_car_weight=0.15,
                                                        vel_div_weight=0.2,
                                                        ), False, True, False),
                             AugmentSetter(ReplaySetter(high_recovery_replays[i], random_boost=False,
                                                        zero_ball_weight=0.8,
                                                        zero_car_weight=0.2,
-                                                       rotate_car_weight=0.3,
+                                                       rotate_car_weight=0.2,
+                                                       backward_car_weight=0.15,
                                                        vel_div_weight=0.2,
                                                        ), False, True, False),
                             HalfFlip(),

diff --git a/worker_recovery.py b/worker_recovery.py
@@ -23,13 +23,10 @@
                         velocity_pb_w=0.01,
                         boost_gain_w=0.25,
                         punish_boost=True,
-                        touch_ball_w=2,
+                        touch_ball_w=3,
                         boost_remain_touch_w=.75,
                         touch_grass_w=-0.03,
                         supersonic_bonus_vpb_w=0.02,
-                        zero_touch_grass_if_ss=True,
-                        turtle_w=-0.04,
-                        final_reward_ball_dist_w=1,
                         )
     frame_skip = Constants_recovery.FRAME_SKIP
     fps = 120 // frame_skip