update examples

Former-commit-id: 369294b31c8a03a1cafcee83eb31a817007d3c49

src/llmtuner/train/sft/metric.py

@@ -2,7 +2,6 @@ from dataclasses import dataclass
 from typing import TYPE_CHECKING, Dict, Sequence, Tuple, Union
 
 import numpy as np
-from transformers.utils.versions import require_version
 
 from ...extras.constants import IGNORE_INDEX
 from ...extras.packages import is_jieba_available, is_nltk_available, is_rouge_available
@@ -33,10 +32,6 @@ class ComputeMetrics:
         r"""
         Uses the model predictions to compute metrics.
         """
-        require_version("jieba", "To fix: pip install jieba")
-        require_version("nltk", "To fix: pip install nltk")
-        require_version("rouge_chinese", "To fix: pip install rouge-chinese")
-
         preds, labels = eval_preds
         score_dict = {"rouge-1": [], "rouge-2": [], "rouge-l": [], "bleu-4": []}