output reforming

              seed=seed)

      else:

        model.fit(train_dataset)

        train_score['user_defined'] = model.evaluate(train_dataset, metric,

                                                     transformers)

        valid_score['user_defined'] = model.evaluate(valid_dataset, metric,

                                                     transformers)

      with open(os.path.join(out_path, 'results.csv'), 'a') as f:

        writer = csv.writer(f)

        for i in valid_score:

          output_line = [

              dataset, str(split), mode, 'valid', i, valid_score[i],

              'time_for_running', time_finish_fitting - time_start_fitting

          ]

          output_line = [dataset, str(split), mode, 'valid', i]

          for task in valid_score[i][0]:

            output_line.extend(

                [task.valid_score[i][0][task], valid_score[i][1][task]])

          output_line.extend(

              ['time_for_running', time_finish_fitting - time_start_fitting])

          writer.writerow(output_line)

datasets = args.dataset_args

cross_valid = args.cross_valid

splitters = ['task']

if len(models) == 0:

  models = ['siamese', 'attn', 'res']

if len(datasets) == 0: