diff --git a/training/eval_history.json b/training/eval_history.json index ed15284b..49c0001d 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -630,5 +630,32 @@ "team_1": 43 }, "win_rate_a": 0.48 + }, + { + "timestamp": "2026-08-15T13:29:47+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260815-0812-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 44, + "wins_b": 30, + "draws": 26, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 23, + "wins_b": 12, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 33, + "team_1": 41 + }, + "win_rate_a": 0.44 } ] diff --git a/training/generation5_state.json b/training/generation5_state.json index 805a4574..3e4736b7 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, @@ -172,6 +172,92 @@ } ], "decision": "fail" + }, + { + "stage_index": 0, + "stage_number": 4, + "stage_name": "handling", + "experiment": "20260815-0812-gen5-s4-handling-retry2", + "attempt": 2, + "telemetry_tail": { + "rollout/air_touch_fraction": 0.0002199999950826168, + "rollout/airborne_fraction": 0.4502894285917282, + "rollout/ep_len_mean": 141.1317198638916, + "rollout/ep_rew_mean": -4.833696460247039, + "rollout/forward_motion_fraction": 0.2373267139494419, + "rollout/goal_rate": 0.5526000013351441, + "rollout/grounded_upright_fraction": 0.051729999739676714, + "rollout/mean_altitude": 4.42696008014679, + "rollout/productive_air_touch_fraction": 7.999999821186065e-05, + "rollout/upright_fraction": 0.2735673332810402, + "rollout/vertical_thrust_mean": 0.08471799899730831 + }, + "telemetry_failures": [ + "rollout/goal_rate=0.5526 < 0.8000", + "rollout/upright_fraction=0.2736 < 0.4500", + "rollout/forward_motion_fraction=0.2373 < 0.2500" + ], + "evaluation_goal_failures": [ + "easy.json: goal_rate=0.740 < 0.800" + ], + "side_balance_failures": [], + "eval": { + "timestamp": "2026-08-15T13:29:47+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260815-0812-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 44, + "wins_b": 30, + "draws": 26, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 23, + "wins_b": 12, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 33, + "team_1": 41 + }, + "win_rate_a": 0.44 + }, + "evals": [ + { + "timestamp": "2026-08-15T13:29:47+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260815-0812-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 44, + "wins_b": 30, + "draws": 26, + "side_results": { + "a_team_0": { + "wins_a": 21, + "wins_b": 18, + "draws": 11 + }, + "a_team_1": { + "wins_a": 23, + "wins_b": 12, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 33, + "team_1": 41 + }, + "win_rate_a": 0.44 + } + ], + "decision": "fail" } ] }