def generate_samples(self, iteration): # generating 50 samples and taking the mean
iter_spacegan = get_spacegan_config(iteration, self.prob_config, self.check_config, self.cond_input, self.target)
# training samples
gan_samples_df = pd.DataFrame(index=range(self.cond_input.shape[0]), columns=self.cond_vars + self.neighbour_list + self.output_vars)
gan_samples_df[self.cond_vars + self.neighbour_list] = self.cond_input
gan_samples_df[self.output_vars] = self.target
for i in range(self.check_config["n_samples"]):
gan_samples_df["sample_" + str(i)] = iter_spacegan.predict(self.cond_input)
return gan_samples_df
# computing metrics
gan_metrics = compute_metrics(target, cond_input, prob_config, check_config,
coord_input, neighbours)
# selecting and sampling gan
for criteria in list(check_config["perf_metrics"].keys()):
# find best config
criteria_info = check_config["pf_metrics_setting"][criteria]
perf_metrics = gan_metrics[criteria_info["metric_level"]]
perf_values = criteria_info["agg_function"](perf_metrics[[criteria]])
best_config = perf_metrics.index[criteria_info["rank_function"](
perf_values)]
# get and set best space gan
best_spacegan = get_spacegan_config(int(best_config), prob_config,
check_config, cond_input, target)
# training samples
gan_samples_df = pd.DataFrame(index=range(cond_input.shape[0]),
columns=cond_vars + neighbour_list +
output_vars)
gan_samples_df[cond_vars + neighbour_list] = cond_input
gan_samples_df[output_vars] = target
for i in range(check_config["n_samples"]):
gan_samples_df["sample_" + str(i)] = best_spacegan.predict(
gan_samples_df[cond_vars + neighbour_list])
# export results
gan_samples_df.to_pickle(model_save_prefix + "grid_" + criteria +
".pkl.gz")
gan_metrics["agg_metrics"].to_pickle(model_save_prefix +
"grid_checkmetrics.pkl.gz")
target, cond_input, spacegan_config.prob_config,
spacegan_config.check_config, coord_input, spacegan_config.neighbours)
# selecting and sampling gan
for criteria in list(spacegan_config.check_config["perf_metrics"].keys()):
# find best config
criteria_info = spacegan_config.check_config["pf_metrics_setting"][
criteria]
perf_metrics = gan_metrics[criteria_info["metric_level"]]
perf_values = criteria_info["agg_function"](perf_metrics[[criteria]])
best_config = perf_metrics.index[criteria_info["rank_function"](
perf_values)]
# get and set best space gan
best_spacegan = spacegan_selection.get_spacegan_config(
int(best_config), spacegan_config.prob_config,
spacegan_config.check_config, cond_input, target)
# training samples
gan_samples_df = pd.DataFrame(index=range(cond_input.shape[0]),
columns=spacegan_config.cond_vars +
neighbour_list +
spacegan_config.output_vars)
gan_samples_df[spacegan_config.cond_vars + neighbour_list] = cond_input
gan_samples_df[spacegan_config.output_vars] = target
for i in range(spacegan_config.check_config["n_samples"]):
gan_samples_df["sample_" + str(i)] = best_spacegan.predict(
gan_samples_df[spacegan_config.cond_vars + neighbour_list])
# export results
gan_samples_df.to_pickle("grid_" + criteria + ".pkl.gz")
spacegan.df_losses.to_pickle("grid_spaceganlosses.pkl.gz")