Problem with total likelihood

[KaisaridiSofia]

In the code in abstract_model we compute :

for k in ("nll_attach", "nll_regul_ind_sum"):
suff_stats[k] = state[k]
suff_stats["nll_tot"] = (
suff_stats["nll_attach"] + suff_stats["nll_regul_ind_sum"]
)

1st issue : At the end of the fit it does not appear in the state, we can not print it. We have the same problem also with :

• "nll_regul_pop_sum" (calculated as : "nll_regul_log_g" + "nll_regul_log_v0" + "nll_regul_betas"
• "nll_regul_all_sum" (calculated as "nll_regul_pop_sum" + nll_regul_ind_sum)

the terms inside the parentheses are well calculated in the end

2nd issue : It's not calculated correctly. It should also add the term "nll_regul_pop_sum"