[BUG] 你好，请问qwen的tokenizer都是slow吗，有没有fast的tokenizer? - Githubissues

QwenLM / Qwen

The official repo of Qwen (通义千问) chat & pretrained large language model proposed by Alibaba Cloud.

Apache License 2.0

13.59k stars 1.11k forks source link

[BUG] 你好，请问qwen的tokenizer都是slow吗，有没有fast的tokenizer? #1006

Closed berooo closed 8 months ago

berooo commented 8 months ago

是否已有关于该错误的issue或讨论？ | Is there an existing issue / discussion for this?

[X] 我已经搜索过已有的issues和讨论 | I have searched the existing issues / discussions

该问题是否在FAQ中有解答？ | Is there an existing answer for this in FAQ?

[X] 我已经搜索过FAQ | I have searched FAQ

当前行为 | Current Behavior

No response

期望行为 | Expected Behavior

No response

复现方法 | Steps To Reproduce

No response

运行环境 | Environment

- OS:
- Python:
- Transformers:
- PyTorch:
- CUDA (`python -c 'import torch; print(torch.version.cuda)'`):

备注 | Anything else?

No response

jklj077 commented 8 months ago

transformers的slow和fast可以理解成，slow是huggingface transformers的实现，fast是huggingface tokenizers的实现。Qwen目前是transformers调用tiktoken，tiktoken的速度高于tokenizers（"fast"）。

想用原生slow/fast的话，文件在这里 https://huggingface.co/Qwen/Qwen-tokenizer