mirror of
https://github.com/langgenius/dify-docs-archived.git
synced 2026-07-25 13:35:35 -04:00
本地社区版部署版本不兼容onnxruntime #84
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @syl51168 on GitHub (Mar 3, 2025).
你们使用的python本本是3.12.0 但是onnxruntime 官方只支持 Python 3.8 - 3.11,而你使用的是 Python 3.12,还怎么玩
@dosubot[bot] commented on GitHub (Mar 3, 2025):
@syl51168 你好!我是Dosu,一个可以帮助你解决 bug、回答问题并成为贡献者的机器人。在等待人类维护者的过程中,我可以帮助你解决问题。
Dify 文档中指定的本地社区版部署使用的是 Python 3.12 [1]。如果 onnxruntime 目前只支持 Python 3.8 到 3.11,您可能需要考虑使用虚拟环境来安装兼容的 Python 版本,或者等待 onnxruntime 更新以支持 Python 3.12。
To continue talking to Dosu, mention @dosu.
Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other
@syl51168 commented on GitHub (Mar 3, 2025):
你们官方本地部署是如何部署的呢?
还是说给的教程就是错的
@dosubot[bot] commented on GitHub (Mar 3, 2025):
官方本地部署的步骤如下:
通过 PyPI 安装 Xinference:
本地启动 Xinference:
这将在本地启动一个工作节点,端点为
http://127.0.0.1:9997。默认情况下,访问仅限于本地机器,但可以通过-H 0.0.0.0配置以允许任何非本地客户端访问。更多详细的配置选项可以使用xinference-local --help查看。访问
http://127.0.0.1:9997,选择所需的模型和规格进行创建和部署。确保模型与您的硬件平台兼容,可以查看 Xinference 内置模型。从部署页面获取模型 UID,这是在 Dify 等应用中连接模型所必需的。
教程是正确的。此外,如果使用 Docker 部署 Dify,确保网络设置允许 Dify 容器访问 Xinference 端点,使用主机机器的 IP 地址而不是 localhost [1][2]。
To continue talking to Dosu, mention @dosu.
Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other
@syl51168 commented on GitHub (Mar 3, 2025):
本地部署,只有中间件是docker的方式跑起来的,api还是直接本地目录部署的