Update tokenization_chatglm.py

https://github.com/huggingface/transformers/blob/8f2b6d5e3dcf40ab0d01f3c8117d1df09e465616/src/transformers/tokenization_utils_base.py#L43

`TensorType` should come from `transformers.utils` instead of `torch`. Otherwise, the tokenizer can't be used alone in an environment without PyTorch installed, where we only need to count tokens.

tokenization_chatglm.py

@@ -3,10 +3,9 @@ import base64
 import os
 import json
 import tiktoken
-from torch import TensorType
 from typing import List, Optional, Union, Dict, Any
 from transformers import PreTrainedTokenizer
-from transformers.utils import logging, PaddingStrategy
+from transformers.utils import logging, PaddingStrategy, TensorType
 from transformers.tokenization_utils_base import EncodedInput, BatchEncoding