feat(*): Log live goal rate to TensorBoard during training

2026-09-13 21:12:03 +00:00 · 2026-07-28 21:33:41 +01:00
parent c998d2271a
commit bca08d266e
4 changed files with 50 additions and 4 deletions
@@ -89,7 +89,11 @@ real behaviour needs tens of millions of steps (hours on the 3090 box).

 Key curves: `rollout/ep_rew_mean` (should trend up), `rollout/ep_len_mean`
 (should trend *down* from 225 as goals end episodes early — 225 action steps
-= the 30s episode timeout).
+= the 30s episode timeout), `rollout/goal_rate` (fraction of recent episodes
+that ended in an actual goal rather than timing out as a draw — the live
+signal for "is the policy actually finishing more episodes by scoring",
+since `ep_rew_mean` mixes that with dense reward-shaping (ball chasing/
+touching) and doesn't isolate it).

 ### Reward/observation tuning