diff --git a/training/eval_history.json b/training/eval_history.json index 0db3e425..5caf7611 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -333,5 +333,32 @@ "team_1": 42 }, "win_rate_a": 0.42 + }, + { + "timestamp": "2026-08-10T07:22:52+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260810-0211-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 39, + "wins_b": 34, + "draws": 27, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 22, + "wins_b": 17, + "draws": 11 + } + }, + "physical_team_wins": { + "team_0": 34, + "team_1": 39 + }, + "win_rate_a": 0.39 } ] diff --git a/training/generation5_state.json b/training/generation5_state.json index ff07c7ed..0a4ba11b 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 +172,91 @@ } ], "decision": "fail" + }, + { + "stage_index": 0, + "stage_number": 4, + "stage_name": "handling", + "experiment": "20260810-0211-gen5-s4-handling-retry2", + "attempt": 2, + "telemetry_tail": { + "rollout/air_touch_fraction": 0.00013999999687075616, + "rollout/airborne_fraction": 0.4427933331727982, + "rollout/ep_len_mean": 141.45631986999513, + "rollout/ep_rew_mean": -0.7604860214497894, + "rollout/forward_motion_fraction": 0.22091919091343878, + "rollout/goal_rate": 0.542920001089573, + "rollout/mean_altitude": 4.399289563179016, + "rollout/productive_air_touch_fraction": 3.999999910593033e-05, + "rollout/upright_fraction": 0.31741619047522546, + "rollout/vertical_thrust_mean": 0.07430099904080387 + }, + "telemetry_failures": [ + "rollout/goal_rate=0.5429 < 0.8000", + "rollout/upright_fraction=0.3174 < 0.4500", + "rollout/forward_motion_fraction=0.2209 < 0.2500" + ], + "evaluation_goal_failures": [ + "easy.json: goal_rate=0.730 < 0.800" + ], + "side_balance_failures": [], + "eval": { + "timestamp": "2026-08-10T07:22:52+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260810-0211-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 39, + "wins_b": 34, + "draws": 27, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 22, + "wins_b": 17, + "draws": 11 + } + }, + "physical_team_wins": { + "team_0": 34, + "team_1": 39 + }, + "win_rate_a": 0.39 + }, + "evals": [ + { + "timestamp": "2026-08-10T07:22:52+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/20260810-0211-gen5-s4-handling-retry2.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/promoted/easy.json", + "seed": 1, + "episodes": 100, + "wins_a": 39, + "wins_b": 34, + "draws": 27, + "side_results": { + "a_team_0": { + "wins_a": 17, + "wins_b": 17, + "draws": 16 + }, + "a_team_1": { + "wins_a": 22, + "wins_b": 17, + "draws": 11 + } + }, + "physical_team_wins": { + "team_0": 34, + "team_1": 39 + }, + "win_rate_a": 0.39 + } + ], + "decision": "fail" } ] }