17 days ago
-
card: the browser player runs iteration 3 seed 2
-
web: web/play.wgr, sha256 b830227b
-
iteration 1 is train_tiny.py at 23e6515 (178 lines): seeds 0-2; card: score 0.4 gfvd + 0.2 logic + 0.2 memory + 0.1 reward + 0.1 done, iteration 1 0.287, 2 0.463, 3 0.562
18 days ago
-
web: web/play.wgr, sha256 f75bfe69
-
iteration 3: an eighth of the windows start a game (0.518 ± 0.010), seeds 0-2; model card
19 days ago
-
v5L checkpoints, seeds 0-2: iteration 1 (train_tiny.py), iteration 2 (train.py@80b1c34), iteration 3 (compact, 5e19333); trained and scored on colab a100
-
iterations 1-3 trained on the v3 data (seeds 0-2, A100) and the 5x model (two seeds)
-
add policy sym_survive
-
add policy opd_1
-
add policy h2_p1
-
add policy walker
-
initial commit
