Fix outdated imports and fix convergence in tabular cfr

[svtuck]

1. Update outdated imports.
2. Fix a bug in tabular cfr that caused it to not converge to the correct cfr for leduc

For leduc this was not converging

       step     EV(avg)
        200    -0.25746
        400    -0.18869
        600    -0.13708
        800    -0.10717
       1000    -0.09344
       1200    -0.07127
       1400    -0.05080
       1600    -0.04214
       1800    -0.03776
       2000    -0.04660

because we were passing the wrong vector to children.

with the fix we get

       step     EV(avg)
      200    -0.15309
      400    -0.11610
      600    -0.10365
      800    -0.09716
     1000    -0.09333
     1200    -0.09076
     1400    -0.08929
     1600    -0.08800
     1800    -0.08710
     2000    -0.08651

which is the right leduc nash ev