tensorflow - 如何在推理阶段使用 TensorFlow 和 I 在抽象文本摘要中的解码序列中使用波束搜索而不是贪婪搜索

Question

def decode_sequence_seq2seq_model_with_just_lstm(
    input_sequence, encoder_model, decoder_model
):
    # Encode the input as state vectors.
    e_out, e_h, e_c = encoder_model.predict(input_sequence)

    # Generate empty target sequence of length 1.
    target_seq = np.zeros((1, 1))

    # Populate the first word of target sequence with the start word.
    target_seq[0, 0] = target_word_index[start_token]

    stop_condition = False
    decoded_sentence = ''

    while not stop_condition:
        output_tokens, h, c = decoder_model.predict(
            [target_seq] + [e_out, e_h, e_c]
        )

        # Sample a token
        sampled_token_index = np.argmax(output_tokens[0, -1, )
        sampled_token = reverse_target_word_index[sampled_token_index]

        if sampled_token != end_token:
            decoded_sentence += ' ' + sampled_token

        # Exit condition: either hit max length or find stop word.
        if (sampled_token == end_token) or (len(decoded_sentence.split()) >= (max_summary_len - 1)):
            stop_condition = True

        # Update the target sequence (of length 1).
        target_seq = np.zeros((1, 1))
        target_seq[0, 0] = sampled_token_index

        # Update internal states
        e_h, e_c = h, c

    return decoded_sentence