diff --git a/training/eval_history.json b/training/eval_history.json index 1aa62109..b9685416 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -252,5 +252,32 @@ "team_1": 46 }, "win_rate_a": 0.43 + }, + { + "timestamp": "2026-08-09T08:36:44+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-0328-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 46, + "wins_b": 29, + "draws": 25, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 11, + "draws": 14 + } + }, + "physical_team_wins": { + "team_0": 32, + "team_1": 43 + }, + "win_rate_a": 0.46 } ] diff --git a/training/generation5_state.json b/training/generation5_state.json index 096b181a..dd863eba 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": "20260809-0328-gen5-s4-handling-retry2", + "attempt": 2, + "telemetry_tail": { + "rollout/air_touch_fraction": 0.0002799999937415123, + "rollout/airborne_fraction": 0.4334809045791626, + "rollout/ep_len_mean": 136.65584008789062, + "rollout/ep_rew_mean": 7.696315350532532, + "rollout/forward_motion_fraction": 0.23557147654891014, + "rollout/goal_rate": 0.5825200036764145, + "rollout/mean_altitude": 4.346557500362397, + "rollout/productive_air_touch_fraction": 3.999999910593033e-05, + "rollout/upright_fraction": 0.2288729049563408, + "rollout/vertical_thrust_mean": 0.06853299860764295 + }, + "telemetry_failures": [ + "rollout/goal_rate=0.5825 < 0.8000", + "rollout/upright_fraction=0.2289 < 0.4500", + "rollout/forward_motion_fraction=0.2356 < 0.2500" + ], + "evaluation_goal_failures": [ + "easy.json: goal_rate=0.750 < 0.800" + ], + "side_balance_failures": [], + "eval": { + "timestamp": "2026-08-09T08:36:44+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-0328-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 46, + "wins_b": 29, + "draws": 25, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 11, + "draws": 14 + } + }, + "physical_team_wins": { + "team_0": 32, + "team_1": 43 + }, + "win_rate_a": 0.46 + }, + "evals": [ + { + "timestamp": "2026-08-09T08:36:44+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-0328-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 46, + "wins_b": 29, + "draws": 25, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 11, + "draws": 14 + } + }, + "physical_team_wins": { + "team_0": 32, + "team_1": 43 + }, + "win_rate_a": 0.46 + } + ], + "decision": "fail" } ] }