checkpoint @ game=2020 steps=2020

training_log.txt

@@ -2318,3 +2318,13 @@
 [2026-06-19 03:38:48] game= 2008 | 1/2-1/2 | moves= 55 | buf=50,000 | steps=2,008 | loss=8.6380 (p=8.4493 v=0.1887) | game=5.6s train=86.0s
 [2026-06-19 03:40:18] game= 2009 | 1/2-1/2 | moves= 49 | buf=50,000 | steps=2,009 | loss=8.6489 (p=8.4493 v=0.1996) | game=4.9s train=85.8s
 [2026-06-19 03:41:49] game= 2010 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=2,010 | loss=8.6281 (p=8.4493 v=0.1788) | game=5.1s train=85.8s
+[2026-06-19 03:43:26] game= 2011 | 1/2-1/2 | moves= 63 | buf=50,000 | steps=2,011 | loss=8.6005 (p=8.4493 v=0.1512) | game=6.3s train=87.8s
+[2026-06-19 03:44:57] game= 2012 | 1/2-1/2 | moves= 50 | buf=50,000 | steps=2,012 | loss=8.6016 (p=8.4493 v=0.1523) | game=5.0s train=86.0s
+[2026-06-19 03:46:29] game= 2013 | 1/2-1/2 | moves= 65 | buf=50,000 | steps=2,013 | loss=8.6033 (p=8.4493 v=0.1540) | game=6.5s train=86.0s
+[2026-06-19 03:48:03] game= 2014 | 1/2-1/2 | moves= 56 | buf=50,000 | steps=2,014 | loss=8.6036 (p=8.4493 v=0.1543) | game=5.6s train=87.9s
+[2026-06-19 03:49:35] game= 2015 | 1/2-1/2 | moves= 65 | buf=50,000 | steps=2,015 | loss=8.5685 (p=8.4493 v=0.1192) | game=6.5s train=86.1s
+[2026-06-19 03:51:13] game= 2016 | 1/2-1/2 | moves= 69 | buf=50,000 | steps=2,016 | loss=8.9529 (p=8.4493 v=0.5036) | game=11.6s train=86.0s
+[2026-06-19 03:52:47] game= 2017 | 1/2-1/2 | moves= 81 | buf=50,000 | steps=2,017 | loss=8.8650 (p=8.4493 v=0.4157) | game=8.0s train=86.2s
+[2026-06-19 03:54:18] game= 2018 | 1/2-1/2 | moves= 48 | buf=50,000 | steps=2,018 | loss=8.8194 (p=8.4493 v=0.3701) | game=4.8s train=86.2s
+[2026-06-19 03:55:53] game= 2019 | 1/2-1/2 | moves= 70 | buf=50,000 | steps=2,019 | loss=8.7985 (p=8.4493 v=0.3492) | game=7.0s train=88.0s
+[2026-06-19 03:57:25] game= 2020 | 1/2-1/2 | moves= 51 | buf=50,000 | steps=2,020 | loss=8.7426 (p=8.4493 v=0.2933) | game=5.1s train=86.3s