846e1fc00dd182b77494ff790d8e0352d756dfc0,t2t_csaky/scripts/benchmark_avg_embedding_sts.py,,main,#,28

Before Change

  output_file_path_fst = os.path.join(
      args.output, "{}-first{}".format(file[0], file[1]))

  output_file_path_snd = os.path.join(
      args.output, "{}-second{}".format(file[0], file[1]))

  split_input_fst, split_input_snd = generate_input_data_for_model(
      args.input, file, args.output)

After Change

        vocab[word] = vocab.get(word, 0) + 1
        word_count += 1

  for word in vocab:
    vocab[word] /= word_count

  dictionary = {}
  with open(args.vocab, "r") as v:
    for line in v:
      line_as_list = line.strip().split()
      if line_as_list[0] in vocab:
        dictionary[line_as_list[0]] = (vocab[line_as_list[0]], numpy.array(
          [float(num) for num in line_as_list[1:]]))
        del vocab[line_as_list[0]]

  del vocab

  create_benchmark(args.input, dictionary)

Instances