A question about placeholder_fulfill function

[Vincent-2017]

`
def placeholder_fulfill(self, prompt: str, length: int, motion_string: str,
text: str):

    seconds = math.floor(length / self.framerate)
    motion_splited = motion_string.split('>')
    token_length = length / self.down_t
    predict_head = int(token_length * self.predict_ratio + 1)
    masked_head = int(token_length * self.inbetween_ratio + 1)
    masked_tail = int(token_length * (1 - self.inbetween_ratio) + 1)
    
    motion_predict_head = '>'.join(
        motion_splited[:predict_head]
    ) + f'><motion_id_{self.m_codebook_size+1}>'
    motion_predict_last = f'<motion_id_{self.m_codebook_size}>' + '>'.join(
        motion_splited[predict_head:])

    motion_masked = '>'.join(
        motion_splited[:masked_head]
    ) + '>' + f'<motion_id_{self.m_codebook_size+2}>' * (
        masked_tail - masked_head) + '>'.join(motion_splited[masked_tail:])

    if random.random() < self.quota_ratio:
        text = f'\"{text}\"'

    prompt = prompt.replace('<Caption_Placeholder>', text).replace(
        '<Motion_Placeholder>',
        motion_string).replace('<Frame_Placeholder>', f'{length}').replace(
            '<Second_Placeholder>', '%.1f' % seconds).replace(
                '<Motion_Placeholder_s1>', motion_predict_head).replace(
                    '<Motion_Placeholder_s2>',
                    motion_predict_last).replace(
                        '<Motion_Placeholder_Masked>', motion_masked)

    return prompt

`

in the code, motion_splited length is 4x for token_length, you use token_length to calculate the mask and then apply the mask in motion_splited, is it correct? I have this question in MotionGPT project.