checkpoint @ game=3590 steps=3590

training_log.txt

@@ -3889,3 +3889,13 @@
 [2026-06-22 12:47:47] game= 3578 | 1/2-1/2 | moves= 57 | buf=50,000 | steps=3,578 | loss=8.7353 (p=8.4493 v=0.2860) | game=5.7s train=87.7s
 [2026-06-22 12:49:23] game= 3579 | 1/2-1/2 | moves= 81 | buf=50,000 | steps=3,579 | loss=8.7005 (p=8.4493 v=0.2512) | game=8.1s train=87.6s
 [2026-06-22 12:50:58] game= 3580 | 1/2-1/2 | moves= 77 | buf=50,000 | steps=3,580 | loss=8.7308 (p=8.4493 v=0.2815) | game=7.7s train=87.6s
+[2026-06-22 12:52:39] game= 3581 | 1/2-1/2 | moves= 67 | buf=50,000 | steps=3,581 | loss=8.6711 (p=8.4493 v=0.2218) | game=6.7s train=90.4s
+[2026-06-22 12:54:13] game= 3582 | 1/2-1/2 | moves= 60 | buf=50,000 | steps=3,582 | loss=8.6788 (p=8.4493 v=0.2295) | game=6.0s train=87.9s
+[2026-06-22 12:55:45] game= 3583 | 1/2-1/2 | moves= 52 | buf=50,000 | steps=3,583 | loss=8.6433 (p=8.4493 v=0.1940) | game=5.2s train=87.7s
+[2026-06-22 12:57:19] game= 3584 | 1/2-1/2 | moves= 62 | buf=50,000 | steps=3,584 | loss=8.6425 (p=8.4493 v=0.1932) | game=6.2s train=87.7s
+[2026-06-22 12:58:53] game= 3585 | 1/2-1/2 | moves= 59 | buf=50,000 | steps=3,585 | loss=8.6223 (p=8.4493 v=0.1730) | game=5.9s train=87.7s
+[2026-06-22 13:00:31] game= 3586 | 1/2-1/2 | moves= 75 | buf=50,000 | steps=3,586 | loss=8.6372 (p=8.4493 v=0.1879) | game=7.6s train=90.2s
+[2026-06-22 13:02:07] game= 3587 | 1/2-1/2 | moves= 81 | buf=50,000 | steps=3,587 | loss=8.6057 (p=8.4493 v=0.1564) | game=8.2s train=87.9s
+[2026-06-22 13:03:42] game= 3588 | 1/2-1/2 | moves= 48 | buf=50,000 | steps=3,588 | loss=9.1449 (p=8.4493 v=0.6956) | game=7.4s train=87.7s
+[2026-06-22 13:05:15] game= 3589 | 1/2-1/2 | moves= 55 | buf=50,000 | steps=3,589 | loss=9.0030 (p=8.4493 v=0.5537) | game=5.5s train=88.0s
+[2026-06-22 13:06:51] game= 3590 | 1/2-1/2 | moves= 81 | buf=50,000 | steps=3,590 | loss=8.9226 (p=8.4493 v=0.4733) | game=8.1s train=87.8s