From 2de3085f13e3e92254b812ba57443fbf7d0cd3fd Mon Sep 17 00:00:00 2001 From: CosmicClash Training Bot Date: Wed, 22 Jul 2026 20:33:07 +0100 Subject: [PATCH] chore(training): curriculum progress after curric-s5-aggression --- training/curriculum_state.json | 22 +++++++++++++++++++--- training/eval_history.json | 10 ++++++++++ 2 files changed, 29 insertions(+), 3 deletions(-) diff --git a/training/curriculum_state.json b/training/curriculum_state.json index 3e6f70f3..33033af9 100644 --- a/training/curriculum_state.json +++ b/training/curriculum_state.json @@ -1,5 +1,5 @@ { - "stage_index": 4, + "stage_index": 5, "attempt": 0, "status": "in_progress", "log": [ @@ -50,7 +50,7 @@ "win_rate_a": 0.26 }, "decision": "fail", - "note": "Original eval (44-19, recorded 2026-07-21T20:46:34) predates the locomotion-mask inference bugfix (8c15c46) and ran with the grounded stage-2 reference unfairly unmasked. Re-run post-fix with --grounded-b reverses the verdict: stage 3's full-3D unmask is a clear regression from stage 2, not an improvement. Not retried via the normal flag-retry mechanism — see stage_index 4 (aggression), which redirects around this branch by resuming from curric-s2-defend directly instead." + "note": "Original eval (44-19, recorded 2026-07-21T20:46:34) predates the locomotion-mask inference bugfix (8c15c46) and ran with the grounded stage-2 reference unfairly unmasked. Re-run post-fix with --grounded-b reverses the verdict: stage 3's full-3D unmask is a clear regression from stage 2, not an improvement. Not retried via the normal flag-retry mechanism \u2014 see stage_index 4 (aggression), which redirects around this branch by resuming from curric-s2-defend directly instead." }, { "stage_index": 3, @@ -67,7 +67,23 @@ "win_rate_a": 0.26 }, "decision": "pass", - "note": "Passes only against its own (already-regressed) predecessor, curric-s3-no_draws. Evaluated directly against grounded curric-s2-defend (2026-07-22T11:40:07), curric-s4-mechanics also loses clearly: 24-57-19. Do not treat this stage's 'pass' as evidence curric-s4-mechanics is the strongest available model overall — see stage 2's note and stage_index 4 (aggression)." + "note": "Passes only against its own (already-regressed) predecessor, curric-s3-no_draws. Evaluated directly against grounded curric-s2-defend (2026-07-22T11:40:07), curric-s4-mechanics also loses clearly: 24-57-19. Do not treat this stage's 'pass' as evidence curric-s4-mechanics is the strongest available model overall \u2014 see stage 2's note and stage_index 4 (aggression)." + }, + { + "stage_index": 4, + "experiment": "curric-s5-aggression", + "attempt": 0, + "eval": { + "timestamp": "2026-07-22T19:33:07+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/curric-s5-aggression.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/curric-s2-defend.json", + "episodes": 100, + "wins_a": 41, + "wins_b": 47, + "draws": 12, + "win_rate_a": 0.41 + }, + "decision": "pass" } ] } diff --git a/training/eval_history.json b/training/eval_history.json index f1ebda75..73dbfa6f 100644 --- a/training/eval_history.json +++ b/training/eval_history.json @@ -88,5 +88,15 @@ "wins_b": 57, "draws": 19, "win_rate_a": 0.24 + }, + { + "timestamp": "2026-07-22T19:33:07+00:00", + "model_a": "/home/jcreek/ai-training/CosmicClash/Game/bots/curric-s5-aggression.json", + "model_b": "/home/jcreek/ai-training/CosmicClash/Game/bots/curric-s2-defend.json", + "episodes": 100, + "wins_a": 41, + "wins_b": 47, + "draws": 12, + "win_rate_a": 0.41 } ]