调用ollama里的模型会话时长只有5分钟[Bug]

[Jerryye17]

💻 系统环境

Other Linux

📦 部署环境

Docker

🌐 浏览器

Chrome

🐛 问题描述

例如LobeChat调用Ollama里的llama3模型，首次提问，ollama会加载llama3模型，一直会话没问题，但一旦不提问后，5分钟模型会被卸载，当再次问问题，又要重新加载llama3模型，对于本地资源不是特别好的情况下，模型加载基本要达到30-60秒，用户体验比较差。
在ollama open webui里，在通用的高级设置里是可以设置保持会话时长的，但LobeChat没有此设置。

🚦 期望结果

能在LobeChat的ollama模型里，设置保持会话时长。

📷 复现步骤

No response

📝 补充信息

No response

[lobehubbot]

👀 @Jerryye17

Thank you for raising an issue. We will investigate into the matter and get back to you as soon as possible.
Please make sure you have given us as much context as possible.
非常感谢您提交 issue。我们会尽快调查此事，并尽快回复您。 请确保您已经提供了尽可能多的背景信息。

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

💻 System environment

Other Linux

📦 Deployment environment

Docker

🌐 Browser

Chrome

🐛 Problem description

For example, LobeChat calls the llama3 model in Ollama. If you ask a question for the first time, ollama will load the llama3 model. There is no problem for the whole session. However, once you do not ask questions, the model will be unloaded in 5 minutes. When you ask questions again, the llama3 model must be reloaded. For local When the resources are not particularly good, the model loading basically takes 30-60 seconds, and the user experience is relatively poor.
In ollama open webui, you can set the duration of the session in the general advanced settings, but LobeChat does not have this setting.

🚦 Expected results

You can set the duration of the session in LobeChat's ollama model.

📷 Steps to reproduce

No response

📝 Supplementary information

No response

[arvinxx]

在 LobeChat 里做可能不一定行，你得自己设定 ollama 的环境变量吧？

https://lobehub.com/zh/docs/self-hosting/examples/ollama#ollama-%E7%8E%AF%E5%A2%83%E5%8F%98%E9%87%8F

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

It may not necessarily work in LobeChat. You have to set ollama's environment variables yourself, right?

https://lobehub.com/en/docs/self-hosting/examples/ollama#ollama-%E7%8E%AF%E5%A2%83%E5%8F%98%E9%87%8F

[Jerryye17]

感谢，部署时加参数 -e OLLAMA_KEEP_ALIVE="5m" 可以控制模型保持会话的时长

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

Thanks, add the parameter -e OLLAMA_KEEP_ALIVE="5m" when deploying to control the length of time the model maintains the session.

[lobehubbot]

✅ @Jerryye17

This issue is closed, If you have any questions, you can comment and reply.
此问题已经关闭。如果您有任何问题，可以留言并回复。