diff --git a/training/eval_history.json b/training/eval_history.json index 455c0eee..244ac764 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -468,5 +468,32 @@ "team_1": 48 }, "win_rate_a": 0.53 + }, + { + "timestamp": "2026-08-12T09:31:39+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260812-0424-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 37, + "wins_b": 29, + "draws": 34, + "side_results": { + "a_team_0": { + "wins_a": 18, + "wins_b": 20, + "draws": 12 + }, + "a_team_1": { + "wins_a": 19, + "wins_b": 9, + "draws": 22 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 39 + }, + "win_rate_a": 0.37 } ] diff --git a/training/generation5_state.json b/training/generation5_state.json index d51ac1c5..c318dcdc 100644 --- a/training/generation5_state.json +++ b/training/generation5_state.json @@ -1,7 +1,7 @@ { "stage_index": 0, "attempt": 2, - "status": "in_progress", + "status": "blocked", "log": [ { "stage_index": 0, @@ -170,6 +170,91 @@ } ], "decision": "fail" + }, + { + "stage_index": 0, + "stage_number": 4, + "stage_name": "handling", + "experiment": "20260812-0424-gen5-s4-handling-retry2", + "attempt": 2, + "telemetry_tail": { + "rollout/air_touch_fraction": 0.00025999999418854715, + "rollout/airborne_fraction": 0.44747342938184737, + "rollout/ep_len_mean": 133.66904002380372, + "rollout/ep_rew_mean": -2.8346366541385652, + "rollout/forward_motion_fraction": 0.22339595240354537, + "rollout/goal_rate": 0.6036400033831596, + "rollout/mean_altitude": 4.436961054801941, + "rollout/productive_air_touch_fraction": 7.999999821186065e-05, + "rollout/upright_fraction": 0.2713429524004459, + "rollout/vertical_thrust_mean": 0.05840199833672028 + }, + "telemetry_failures": [ + "rollout/goal_rate=0.6036 < 0.8000", + "rollout/upright_fraction=0.2713 < 0.4500", + "rollout/forward_motion_fraction=0.2234 < 0.2500" + ], + "evaluation_goal_failures": [ + "easy.json: goal_rate=0.660 < 0.800" + ], + "side_balance_failures": [], + "eval": { + "timestamp": "2026-08-12T09:31:39+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260812-0424-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 37, + "wins_b": 29, + "draws": 34, + "side_results": { + "a_team_0": { + "wins_a": 18, + "wins_b": 20, + "draws": 12 + }, + "a_team_1": { + "wins_a": 19, + "wins_b": 9, + "draws": 22 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 39 + }, + "win_rate_a": 0.37 + }, + "evals": [ + { + "timestamp": "2026-08-12T09:31:39+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260812-0424-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 37, + "wins_b": 29, + "draws": 34, + "side_results": { + "a_team_0": { + "wins_a": 18, + "wins_b": 20, + "draws": 12 + }, + "a_team_1": { + "wins_a": 19, + "wins_b": 9, + "draws": 22 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 39 + }, + "win_rate_a": 0.37 + } + ], + "decision": "fail" } ] }