* Setting all random seeds to  0 *
Loading model: out_models_control/net_fs_vanilla_rep3_best.pt on cpu
 Loading model that has completed (or started) 1 of 50 epochs
  test episode_type: human_vanilla
  batch size: 1
  max eval length: 10
  number of steps: 1
  best val loss achieved: 1.7322

BIML specs:
 nparams= 1392263
 nlayers_encoder= 3
 nlayers_decoder= 3
 nhead= 8
 hidden_size= 128
 dim_feedforward= 512
 act_feedforward= gelu
 dropout= 0.1
 

Fitting for the best value of p_lapse use log-like...
  Each value is replicated across 2 random runs/permutations
 p_lapse 0.01 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1721.3053 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7564
 p_lapse 0.02 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1717.4741 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7525
 p_lapse 0.03 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1713.7475 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7487
 p_lapse 0.04 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1710.121 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.745
 p_lapse 0.05 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1706.5904 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7414
 p_lapse 0.06 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1703.1516 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7379
 p_lapse 0.07 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1699.801 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7345
 p_lapse 0.08 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1696.535 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7312
 p_lapse 0.09 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1693.3506 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7279
 p_lapse 0.1 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1690.2447 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.7247
 p_lapse 0.2 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1662.9511 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6969
 p_lapse 0.3 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1641.2025 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6747
 p_lapse 0.4 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1623.7253 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6569
 p_lapse 0.5 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1609.7064 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6426
 p_lapse 0.6 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1598.5998 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6312
 p_lapse 0.7 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1590.0273 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6225
 p_lapse 0.8 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1583.724 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.616
 p_lapse 0.9 :
* Setting all random seeds to  0 *
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  skipped DAX thrice after 2 -> 2 undefined_action undefined_action undefined_action
  skipped 2 after 3 surround DAX -> 3 undefined_action 3 2
  loglike: M = -1579.5055 (SD= 0.0 , Nrep= 2 ) for 980 symbol predictions
    ave LL:  -1.6117
* BEST FIT * p_lapse= 0.9 with loglike score of -1579.5055
