diff --git a/xtuner/dataset/internvl_dataset.py b/xtuner/dataset/internvl_dataset.py index 82904ae87..9f2bcab2b 100644 --- a/xtuner/dataset/internvl_dataset.py +++ b/xtuner/dataset/internvl_dataset.py @@ -400,10 +400,11 @@ def get_inputid_labels(self, conversations, image_token_str) -> dict: labels += copy.deepcopy(output_encode) if len(input_ids) > self.max_length: - input_ids = input_ids[:self.max_length] - labels = labels[:self.max_length] print_log( f'Warning: input_ids length({len(input_ids)}) ' f'is longer than max_length, cut to {self.max_length}', logger='current') + input_ids = input_ids[:self.max_length] + labels = labels[:self.max_length] + return {'input_ids': input_ids, 'labels': labels}