# get_tokenizer

> **Function** in `langchain_core`

📖 [View in docs](https://reference.langchain.com/python/langchain-core/language_models/base/get_tokenizer)

Get a GPT-2 tokenizer instance.

This function is cached to avoid re-loading the tokenizer every time it is called.

## Signature

```python
get_tokenizer() -> Any
```

## Returns

`Any`

The GPT-2 tokenizer instance.

---

[View source on GitHub](https://github.com/langchain-ai/langchain/blob/1c3a4186cf2ba4f28face59118ac7786de009f91/libs/core/langchain_core/language_models/base.py#L77)