from scipy.stats import multivariate_normal from tf_util.families import family_from_str from efn_util import model_opt_hps import os, sys os.chdir('../') exp_fam = str(sys.argv[1]) D = int(sys.argv[2]) give_inverse_hint = int(sys.argv[3]) == 1 random_seed = int(sys.argv[4]) dir_str = str(sys.argv[5]) profile = True TIF_flow_type, nlayers, scale_layer, lr_order = model_opt_hps(exp_fam, D) nlayers = 30 lr_order = -3 flow_dict = {'latent_dynamics':None, \ 'scale_layer':False, \ 'TIF_flow_type':TIF_flow_type, \ 'repeats':nlayers} fam_class = family_from_str(exp_fam) family = fam_class(D) family.load_data() family.select_train_test_sets(500) param_net_input_type = 'eta'
import numpy as np from scipy.stats import multivariate_normal from families import family_from_str from efn_util import model_opt_hps import os, sys os.chdir('../') exp_fam = str(sys.argv[1]) D = int(sys.argv[2]) nlayers = int(sys.argv[3]) give_inverse_hint = int(sys.argv[4]) == 1 random_seed = int(sys.argv[5]) dir_str = str(sys.argv[6]) TIF_flow_type, _, lr_order = model_opt_hps(exp_fam, D) flow_dict = {'latent_dynamics':None, \ 'TIF_flow_type':TIF_flow_type, \ 'repeats':nlayers} fam_class = family_from_str(exp_fam) family = fam_class(D) param_net_input_type = 'eta' cost_type = 'KL' K_eta = 100 M_eta = 1000 stochastic_eta = True dist_seed = 0 max_iters = 1000000