[Bug] API请求超时:无法完成html重构(啥活没产,就干了我input 43万个tokens) #829
Replies: 5 comments
-
|
我也是,问了个问题,350万token没了 |
Beta Was this translation helpful? Give feedback.
-
|
+1 10个request 123k的input 顶不住 |
Beta Was this translation helpful? Give feedback.
-
|
+1, it costs me 5.8M tokens for constant error and retry f**king qwen, now 8.7M tokens ??? |
Beta Was this translation helpful? Give feedback.
-
|
maxRetries=3 Streaming request timeout after 1215s Is prompt re-submitted each retry? is maxRetries=2 is double GPU work requested ? "300s" base limit... where is it configured? ( qwen on ubuntu -> ollama-serve on mac ) qwen or ollama or OS (regarding "OLS" /proc/sys/net/ipv4/tcp_keepalive_time on linux and net.inet.tcp.keepidle on MacOS are both 7200 ( 2 hours ) so I don't think it is at the TCP level) ? |
Beta Was this translation helpful? Give feedback.
-
|
If I set maxRetries=-1 , then this error goes away; but the problem surfaces in the model provider logs after 300 seconds:
where can this 300 second timeout be increased? |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
标题
[Bug] API请求超时:无法完成html重构(啥活没产,就干了我input 43万个tokens)描述
在尝试根据新设计规范重构
example.html时,遇到了持续性的API请求超时问题。每次尝试均在308秒后失败,尽管已成功读取文件内容,但无法完成重构操作。重现步骤
example.html文件内容(约[4000]行代码)✦ 按照新设计规范重构example.html,保留JavaScript功能✔ ReadFile example.html)错误日志
Beta Was this translation helpful? Give feedback.
All reactions