about labels and decoder_attention_mask

[gitfor20]

target_text_encoding = self.tokenizer(
data_row["target_text"],
max_length=self.target_max_token_len,
padding="max_length",
truncation=True,
return_attention_mask=True,
add_special_tokens=True,
return_tensors="pt",
)

    labels = target_text_encoding["input_ids"]
    labels[
        labels == 0
    ] = -100  # to make sure we have correct labels for T5 text generation

    return dict(
        source_text_input_ids=source_text_encoding["input_ids"].flatten(),
        source_text_attention_mask=source_text_encoding["attention_mask"].flatten(),
        labels=labels.flatten(),
        labels_attention_mask=target_text_encoding["attention_mask"].flatten(),
    )

as i know, the decoder_input_ids is default to be got by shifting labels, but at these codes, the decoder_attention_mask is matched to labels. so i think the decoder_input_ids prepared by models will not be matched to decoder_attention_mask.
is it a bug or my understanding is wrong?