DFA + residual-branch penalty test: seed=123, lam=0.0001 ep 0: ||h_L||=8.791e+00 ||g_2||=9.848e-04 acc=0.0957 ep 1: ||h_L||=2.827e+03 ||g_2||=3.043e-06 eval_acc=0.3262 test_acc=0.3037 ep 10: ||h_L||=1.476e+04 ||g_2||=9.239e-07 eval_acc=0.3379 test_acc=0.3408 ep 20: ||h_L||=1.954e+04 ||g_2||=7.765e-07 eval_acc=0.3604 test_acc=0.3606 ep 30: ||h_L||=2.057e+04 ||g_2||=7.580e-07 eval_acc=0.3682 test_acc=0.3622 FINAL test acc: 0.3622 Compare to: DFA-vanilla (3-seed mean): 0.308 DFA-shallow (3-seed mean): 0.349 DFA-frozen (3-seed mean): 0.349 BP-trainable (3-seed mean): 0.609 Saved results/dfa_pen_short/dfa_pen_lam0.0001_s123.json Saved results/dfa_pen_short/dfa_pen_lam0.0001_s123.pt