checkpoint @ game=3640 steps=3640

training_log.txt

@@ -3939,3 +3939,13 @@
 [2026-06-23 00:33:25] game= 3628 | 1/2-1/2 | moves= 62 | buf=34,921 | steps=3,628 | loss=8.7350 (p=8.4493 v=0.2857) | game=6.4s train=85.8s
 [2026-06-23 00:34:57] game= 3629 | 1/2-1/2 | moves= 53 | buf=34,974 | steps=3,629 | loss=8.7020 (p=8.4493 v=0.2528) | game=5.7s train=86.1s
 [2026-06-23 00:36:29] game= 3630 | 1/2-1/2 | moves= 60 | buf=35,034 | steps=3,630 | loss=8.6581 (p=8.4493 v=0.2088) | game=6.0s train=86.5s
+[2026-06-23 00:38:05] game= 3631 | 1/2-1/2 | moves= 50 | buf=35,084 | steps=3,631 | loss=8.6252 (p=8.4493 v=0.1759) | game=5.3s train=85.9s
+[2026-06-23 00:39:36] game= 3632 | 1-0 | moves= 41 | buf=35,125 | steps=3,632 | loss=8.6156 (p=8.4493 v=0.1663) | game=4.4s train=86.2s
+[2026-06-23 00:41:11] game= 3633 | 1/2-1/2 | moves= 79 | buf=35,204 | steps=3,633 | loss=8.6260 (p=8.4493 v=0.1767) | game=8.7s train=86.2s
+[2026-06-23 00:42:42] game= 3634 | 1/2-1/2 | moves= 56 | buf=35,260 | steps=3,634 | loss=8.6251 (p=8.4493 v=0.1758) | game=5.9s train=86.0s
+[2026-06-23 00:44:16] game= 3635 | 1/2-1/2 | moves= 72 | buf=35,332 | steps=3,635 | loss=8.5950 (p=8.4493 v=0.1457) | game=7.6s train=86.4s
+[2026-06-23 00:45:49] game= 3636 | 1/2-1/2 | moves= 52 | buf=35,384 | steps=3,636 | loss=8.5933 (p=8.4493 v=0.1440) | game=5.6s train=86.8s
+[2026-06-23 00:47:25] game= 3637 | 1/2-1/2 | moves= 65 | buf=50,000 | steps=3,637 | loss=9.1006 (p=8.4493 v=0.6513) | game=9.9s train=85.9s
+[2026-06-23 00:48:56] game= 3638 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=3,638 | loss=9.0996 (p=8.4493 v=0.6503) | game=5.4s train=86.0s
+[2026-06-23 00:50:29] game= 3639 | 1/2-1/2 | moves= 55 | buf=50,000 | steps=3,639 | loss=8.9368 (p=8.4493 v=0.4875) | game=5.9s train=86.7s
+[2026-06-23 00:52:01] game= 3640 | 1/2-1/2 | moves= 57 | buf=50,000 | steps=3,640 | loss=8.8722 (p=8.4493 v=0.4229) | game=5.9s train=85.9s