Skip to content

Commit 930604d

Browse files
committed
docs: update llm token rate limit usage
1 parent 311ca30 commit 930604d

File tree

2 files changed

+3
-1
lines changed

2 files changed

+3
-1
lines changed

core/llm_token_ratelimit/README_en.md

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -185,6 +185,7 @@ Currently, it supports non-intrusive integration of Langchaingo and Eino framewo
185185
186186
#### Notes
187187
188+
- Since only input tokens can be predicted at present, **it is recommended to use PETA for rate limiting specifically targeting input tokens**
188189
- PETA uses tiktoken to estimate input token consumption but requires downloading or preconfiguring the `Byte Pair Encoding (BPE)` dictionary
189190
- Online mode
190191
- tiktoken needs to download encoding files online for the first use

core/llm_token_ratelimit/README_zh.md

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -178,13 +178,14 @@ sentinel:
178178
```
179179
#### LLM框架适配
180180
181-
目前支持Langchaingo和Eino框架无侵入式接入Sentinel提供的Token限流能力,主要适用于文本生成方面,使用方法详见
181+
目前支持Langchaingo和Eino框架无侵入式接入Sentinel提供的Token限流能力,主要适用于文本生成方面,使用方法详见
182182
183183
- pkg/adapters/langchaingo/wrapper.go
184184
- pkg/adapters/eino/wrapper.go
185185
186186
#### 注意事项
187187
188+
- 由于目前仅可预知input tokens,所以**建议使用PETA专对于input tokens进行限流**
188189
- PETA使用tiktoken预估输入消耗token数,但是需要下载或预先配置`字节对编码(Byte Pair Encoding,BPE)`字典
189190
- 在线模式
190191
- 首次使用时,tiktoken需要联网下载编码文件

0 commit comments

Comments
 (0)