palette = sns.color_palette("bright", 2) sns.scatterplot(DA_embedded[:, 0], DA_embedded[:, 1], hue=y_DA_list, legend='full', palette=palette) plt.savefig(figure_folder + '/DA.png') plt.clf() sns.set(rc={'figure.figsize': (11.7, 8.27)}) palette = sns.color_palette("bright", 2) sns.scatterplot(TO_embedded[:, 0], TO_embedded[:, 1], hue=y_TO_list, legend='full', palette=palette, markers='s') plt.savefig(figure_folder + '/TO.png') from helper_function import plot_feature_dist, plot_feature_pair_dist plot_feature_dist(figure_folder + '/SO.png', SO_features, yt_test, 'SO') plot_feature_dist(figure_folder + '/DA.png', DA_features, yt_test, 'DA') plot_feature_dist(figure_folder + '/TO.png', target_features, yt_test, 'TO') plot_feature_pair_dist(figure_folder + '/DA_TO.png', DA_features, target_features, yt_test, yt_test, label=['DA', 'TO'])
test_auc_list, val_auc_list, src_test_list, DA_model_folder + '/AUC_src_{}.png'.format(DA_model_name)) plot_auc_iterations( test_auc_list, val_auc_list, DA_model_folder + '/AUC_Final_{}.png'.format(DA_model_name)) if best_val_auc < val_target_AUC: best_val_auc = val_target_AUC target_saver.save(sess, DA_model_folder + '/target_best') np.savetxt(os.path.join(DA_model_folder, 'test_stat.txt'), test_target_stat) np.savetxt(os.path.join(DA_model_folder, 'test_best_auc.txt'), [test_target_AUC]) print_red('Update best:' + DA_model_folder) if iteration % 1000 == 0: indices = np.random.randint(0, Xs_tst.shape[0], 100) source_feat = h_src.eval(session=sess, feed_dict={ xs: Xs_tst[indices, ], g_training: False }) target_feat = h_trg.eval(session=sess, feed_dict={ xt: Xt_tst[indices, ], g_training: False }) plot_feature_pair_dist( DA_model_folder + '/feat_{}_iter_{}.png'.format(DA_model_name, iteration), np.squeeze(source_feat), np.squeeze(target_feat), ys_tst[indices], yt_tst[indices], ['source', 'target']) gc.collect()
fig_size=(10, 10)) # save models if iteration % 100 == 0: target_saver.save(sess, DA_model_folder + '/target', global_step=iteration) if best_val_auc < val_target_AUC: best_val_auc = val_target_AUC target_saver.save(sess, DA_model_folder + '/target_best') np.savetxt(os.path.join(DA_model_folder, 'test_stat.txt'), test_target_stat) np.savetxt(os.path.join(DA_model_folder, 'test_best_auc.txt'), [test_target_AUC]) print_red('Update best:' + DA_model_folder) if iteration % 10000 == 0: source_feat = h_src.eval(session=sess, feed_dict={ xs: Xs_tst, is_training: False, dis_training: False }) target_feat = h_trg.eval(session=sess, feed_dict={ xt: Xt_tst, is_training: False, dis_training: False }) plot_feature_pair_dist( DA_model_folder + '/feat_{}.png'.format(DA_model_name), np.squeeze(source_feat), np.squeeze(target_feat), ys_tst, yt_tst, ['source', 'target'])