diff --git a/training/eval_history.json b/training/eval_history.json index 6cd4e34b..0db3e425 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -306,5 +306,32 @@ "team_1": 42 }, "win_rate_a": 0.46 + }, + { + "timestamp": "2026-08-10T01:11:41+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-1955-gen5-s4-handling-retry1.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 42, + "wins_b": 27, + "draws": 31, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 10, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 42 + }, + "win_rate_a": 0.42 } ] diff --git a/training/generation5_state.json b/training/generation5_state.json index 5fc0f238..ff07c7ed 100644 --- a/training/generation5_state.json +++ b/training/generation5_state.json @@ -1,6 +1,6 @@ { "stage_index": 0, - "attempt": 1, + "attempt": 2, "status": "in_progress", "log": [ { @@ -87,6 +87,91 @@ } ], "decision": "fail" + }, + { + "stage_index": 0, + "stage_number": 4, + "stage_name": "handling", + "experiment": "20260809-1955-gen5-s4-handling-retry1", + "attempt": 1, + "telemetry_tail": { + "rollout/air_touch_fraction": 0.00019999999552965163, + "rollout/airborne_fraction": 0.45090347522497176, + "rollout/ep_len_mean": 145.11324035644532, + "rollout/ep_rew_mean": -1.5461844632700086, + "rollout/forward_motion_fraction": 0.21509128576517106, + "rollout/goal_rate": 0.5097999980449677, + "rollout/mean_altitude": 4.43802117395401, + "rollout/productive_air_touch_fraction": 9.999999776482581e-05, + "rollout/upright_fraction": 0.3002223813533783, + "rollout/vertical_thrust_mean": 0.047706998658077794 + }, + "telemetry_failures": [ + "rollout/goal_rate=0.5098 < 0.8000", + "rollout/upright_fraction=0.3002 < 0.4500", + "rollout/forward_motion_fraction=0.2151 < 0.2500" + ], + "evaluation_goal_failures": [ + "easy.json: goal_rate=0.690 < 0.800" + ], + "side_balance_failures": [], + "eval": { + "timestamp": "2026-08-10T01:11:41+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-1955-gen5-s4-handling-retry1.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 42, + "wins_b": 27, + "draws": 31, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 10, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 42 + }, + "win_rate_a": 0.42 + }, + "evals": [ + { + "timestamp": "2026-08-10T01:11:41+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260809-1955-gen5-s4-handling-retry1.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 42, + "wins_b": 27, + "draws": 31, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 25, + "wins_b": 10, + "draws": 15 + } + }, + "physical_team_wins": { + "team_0": 27, + "team_1": 42 + }, + "win_rate_a": 0.42 + } + ], + "decision": "fail" } ] }