1panel-dev / maxkb Goto Github PK
View Code? Open in Web Editor NEW🚀 基于 LLM 大语言模型的知识库问答系统。开箱即用、模型中立、灵活编排,支持快速嵌入到第三方业务系统。
Home Page: https://maxkb.cn/
License: GNU General Public License v3.0
🚀 基于 LLM 大语言模型的知识库问答系统。开箱即用、模型中立、灵活编排,支持快速嵌入到第三方业务系统。
Home Page: https://maxkb.cn/
License: GNU General Public License v3.0
latest
第一次尝试部署 maxkb,我用的是一台4G内存的云服务器,发起启动容器以后服务器不响应,发现内存占用率在猛增。
希望开发者能在文档中添加一下运行环境的硬件需求信息。
No response
1.0
倒逼1panel支持gpu
1panel支持gpu;
maxkb支持调用gpu
No response
No response
1.0
添加分段时无法选择问题
添加分段
可以选择问题
No response
No response
No response
v1.0.0
各位maxkb的工程师,你们好,有个段落绑定的问题。在新创建的段落页面立刻关联问题的话,填写的问题会无法绑定。
在新创建的段落页面立刻关联问题的话,填写的问题会无法绑定。
创建段落和问题绑定可以同时进行
No response
No response
No response
v1.0.0 (build at 2024-04-10T19:09, commit: cbaf704)
参考图片
No response
No response
No response
1.0.0
建议支持指定本地embedding模型
No response
No response
1.0
希望文档可以支持从一个项目迁移到另外一个项目,方便批量整理数据
增加批量迁移按钮,迁移的时候支持选择有权限的项目
No response
v0.9.1
很多文档是采用Markdown编写的,通过直接读取对应的Git仓库,可以直接跳过文档渲染器带来的影响,获取干净的信息。
(最好能够支持Gitea就好了)
注册成gitea/github应用,然后借助webhook触发自动同步更新信息
No response
v1.0.0
如题
如题
No response
v1.0.0
如题。
No response
No response
No response
v1.0.1
应用命中测试报没有权限访问
应用命中测试报没有权限访问
No response
No response
No response
1,0
文档全量更新无法使用,报错
进入文档,点击批量删除
可以批量删除
No response
No response
v0.9.1
如题。
No response
No response
v0.9.0 (build at 2024-03-21T18:39, commit: 96a0e4b)
现在只支持添加 Azure OpenAI 的 API,以及基础模型
只能选 gpt-3.5-turbo
系列和 gpt-4-0613
。
建议增加 OpenAI 格式的自定义模型,并且将基础模型
改为自由输入而不是下拉框。毕竟llama.cpp
、vllm
等开源项目更广泛地支持的是 OpenAI 的 API 格式。
Reference: platform.openai.com
No response
No response
V1.0.0
应用的logo支持自定义上传
No response
No response
1.0
fastapi 即可
No response
No response
v0.9.1
新增用量统计功能,针对每个应用支持统计提问用户数据、新增提问用户数、提问数、 消费Tokens数和用户满意度。
No response
No response
v0.9.1
新增问题库,支持把用户的问题梳理到问题库,关联文档中的分段,从而更能精确匹配到相关知识点。
No response
No response
v0.9.1
【应用】概览中重新生成公开访问链接会导致嵌入第三方脚本失效,增加确认提示。
No response
No response
V1.0.0
希望能限制ai模型只回答知识库中的内容。
No response
No response
0.9.1
支持多语言界面的建议
尊敬的MaxKB开发团队,
您好!我是一名热心的MaxKB用户,非常欣赏MaxKB在知识管理和分享方面的卓越性能。我想提出一个关于软件多语言支持的建议。随着MaxKB用户群体的国际化,增加多语言支持能够让更多的用户更容易地使用MaxKB,促进其在全球范围内的普及和增长。
多语言界面将允许用户选择他们首选的语言,这不仅可以提升用户体验,也能增强MaxKB的可用性。此举也是对多元文化和包容性的重要承诺。
采用vue-i18n库来构建一个本地化的框架。借助大语言模型,实现语言翻译,可以方便快速的开发出多语言的界面
我相信,通过集体的努力,MaxKB可以更好地服务全球用户。期待您的反馈,并愿意参与到实施过程中来。
感谢您的考虑!
版本:v0.9.0
需求:
在系统安装成功后,默认所有用户共享相同的账户密码。
为提升安全性,可以考虑引入强制修改密码功能。
0.9.1
因服务器资源有限,希望PostgreSQL / pgvector能够被重用,能很好的负载均衡,因此期望MaxKB在部署时能够通过环境变量设置指定数据库服务器,来启动服务。
通过环境变量设置指定数据库服务器,来启动服务。
No response
No response
1.0.0
连接的Ollama,每次对话都是回复 抱歉,没有查找到相关内容,请重新描述您的问题或提供更多信息。
这个是maxkb的日志
/opt/py3/lib/python3.11/site-packages/django/db/models/base.py:364: RuntimeWarning: Model 'dataset.dynamics' was already registered. Reloading models is not advised as it can lead to inconsistencies, most notably with related models.
new_class._meta.apps.register_model(new_class._meta.app_label, new_class)
无
No response
No response
No response
1.0
抛开事实和网速不提
No response
No response
1.0.1
本地上传的文档内带的图片能同步到 maxkb 里
本地上传的文档内带的图片能同步到 maxkb 里
No response
v0.9.1
建议可以设置是否显示回答时下面显示引用的知识库的内容,这样更准确的了解回答的内容!
No response
No response
lastest
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
docker run
start up
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
127.0.0.1:5432 - no response
waiting for postgres
main
你好 MaxKB社区,我刚刚阅读了下代码。在 RAG 方面,社区采用多轮识别用户意图,转化成一个新问题,然后发起一个相似搜索,然后一并提交给 LLM ,在保证问题的准确。
仅仅只是相似搜索,其实准确率不太行。可以做下混合搜索,多路的召回方式。比如 关键字 + 相似搜索的混合搜索,或者语义搜索 + 相似搜索的混合搜索。
opensearch 实现了向量化,而且本身也是支持语义搜索的,比如 fuzzy 查询,这两者可以直接配合在一起。
或者其他向量数据库的所有数据组成一个大行,然后自己使用 BM25 全文检索,获取到当行的 id ,反查出来的原文,在一并提交过。
No response
v1.0.0
建议增加分离部署,不是必须使用all-in-one的模式,增加更多灵活性。尤其是vactorDB。
No response
No response
No response
0.9
无论是word转txt,还是新建txt文档,上传后都不能识别,段落是0,字符数也是0,实在无语。另外,居然不支持docx和pdf文件,局限性太大了吧
docx转txt后无法识别
可以正确识别txt文件
2024-03-22 12:56:13 [listener_manage INFO] 开始--->向量化文档:81e41878-e808-11ee-8dfc-0242ac120003
2024-03-22 12:56:13 [basehttp INFO] "POST /api/dataset/04c51c78-e805-11ee-b157-0242ac120003/document/_bach HTTP/1.1" 200 393
2024-03-22 12:56:13 [listener_manage INFO] 结束--->向量化文档:81e41878-e808-11ee-8dfc-0242ac120003
2024-03-22 12:56:14 [basehttp INFO] "GET /ui/assets/icon_document-2fa30876.svg HTTP/1.1" 200 1376
2024-03-22 12:56:14 [basehttp INFO] "GET /api/dataset HTTP/1.1" 200 366
2024-03-22 12:56:14 [basehttp INFO] "GET /api/dataset/04c51c78-e805-11ee-b157-0242ac120003/document/1/10 HTTP/1.1" 200 444
2024-03-22 12:56:14 [basehttp INFO] "GET /api/dataset/04c51c78-e805-11ee-b157-0242ac120003 HTTP/1.1" 200 391
2024-03-22 12:56:14 [basehttp INFO] - Broken pipe from ('222.90.143.25', 35901)
2024-03-22 12:56:14 [basehttp INFO] "GET /ui/assets/icon_document-2fa30876.svg HTTP/1.1" 200 1376
2024-03-22 12:56:15 [basehttp INFO] - Broken pipe from ('222.90.143.25', 36208)
2024-03-22 12:56:16 [basehttp INFO] "GET /ui/assets/index-3d9325d8.js HTTP/1.1" 200 5566
2024-03-22 12:56:16 [basehttp INFO] "GET /ui/assets/index-0432a3d8.css HTTP/1.1" 200 2044
2024-03-22 12:56:17 [basehttp INFO] "GET /ui/assets/icon_document-2fa30876.svg HTTP/1.1" 200 1376
2024-03-22 12:56:17 [basehttp INFO] - Broken pipe from ('222.90.143.25', 36209)
2024-03-22 12:56:20 [basehttp INFO] "GET /api/dataset/04c51c78-e805-11ee-b157-0242ac120003/hit_test?query_text=%E5%AE%89%E5%85%A8&similarity=0.6&top_number=5 HTTP/1.1" 200 52
2024-03-22 12:56:20 [basehttp INFO] "GET /ui/assets/user-icon-c413d294.svg HTTP/1.1" 304 0
2024-03-22 12:56:56 [log WARNING] Unauthorized: /api/dataset/04c51c78-e805-11ee-b157-0242ac120003/document/_bach
2024-03-22 12:56:56 [basehttp WARNING] "HEAD /api/dataset/04c51c78-e805-11ee-b157-0242ac120003/document/_bach HTTP/1.1" 401 86
2024-03-22 12:57:35 [basehttp INFO] "GET /api/user_manage/1/20 HTTP/1.1" 200 369
2024-03-22 12:58:59.115 CST [67] LOG: checkpoint starting: time
2024-03-22 12:59:00.205 CST [67] LOG: checkpoint complete: wrote 11 buffers (0.1%); 0 WAL file(s) added, 0 removed, 0 recycled; write=1.010 s, sync=0.034 s, total=1.090 s; sync files=10, longest=0.028 s, average=0.004 s; distance=1 kB, estimate=837 kB
No response
版本号:v0.9.0 (build at 2024-03-21T18:39, commit: 96a0e4b)
在新增应用时,配置好应用内的细节后,在右侧的测试窗口对话时提示报错。
不报错,返回信息。
waiting for postgres
2024-03-24 19:04:56.830 CST [141] FATAL: the database system is starting up
127.0.0.1:5432 - rejecting connections
waiting for postgres
2024-03-24 19:04:57.845 CST [144] FATAL: the database system is starting up
127.0.0.1:5432 - rejecting connections
waiting for postgres
2024-03-24 19:04:58.860 CST [147] FATAL: the database system is starting up
127.0.0.1:5432 - rejecting connections
2024-03-24 19:04:59.005 CST [31] LOG: database system was not properly shut down; automatic recovery in progress
2024-03-24 19:04:59.051 CST [31] LOG: redo starts at 0/1E75BC0
2024-03-24 19:04:59.051 CST [31] LOG: invalid record length at 0/1E75CA8: wanted 24, got 0
2024-03-24 19:04:59.051 CST [31] LOG: redo done at 0/1E75C70 system usage: CPU: user: 0.00 s, system: 0.00 s, elapsed: 0.00 s
2024-03-24 19:04:59.247 CST [29] LOG: checkpoint starting: end-of-recovery immediate wait
2024-03-24 19:04:59.566 CST [29] LOG: checkpoint complete: wrote 3 buffers (0.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.120 s, sync=0.031 s, total=0.366 s; sync files=2, longest=0.016 s, average=0.016 s; distance=0 kB, estimate=0 kB
2024-03-24 19:04:59.624 CST [7] LOG: database system is ready to accept connections
waiting for postgres
127.0.0.1:5432 - accepting connections
Operations to perform:
Apply all migrations: application, contenttypes, dataset, django_apscheduler, embedding, setting, users
Running migrations:
No migrations to apply.
Your models in app(s): 'application' have changes that are not yet reflected in a migration, and so won't be applied.
Run 'manage.py makemigrations' to make new migrations, and then re-run 'manage.py migrate' to apply them.
Operations to perform:
Apply all migrations: application, contenttypes, dataset, django_apscheduler, embedding, setting, users
Running migrations:
No migrations to apply.
Your models in app(s): 'application' have changes that are not yet reflected in a migration, and so won't be applied.
Run 'manage.py makemigrations' to make new migrations, and then re-run 'manage.py migrate' to apply them.
2024-03-24 19:09:59.593 CST [29] LOG: checkpoint starting: time
2024-03-24 19:10:00.273 CST [29] LOG: checkpoint complete: wrote 4 buffers (0.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.333 s, sync=0.044 s, total=0.681 s; sync files=4, longest=0.029 s, average=0.011 s; distance=3 kB, estimate=3 kB
2024-03-24 19:40:08 [basehttp INFO] "GET /api/provider/model_type_list?provider=model_ollama_provider HTTP/1.1" 200 109
2024-03-24 19:40:08 [basehttp INFO] "GET /api/provider/model_form?provider=model_ollama_provider&model_type=LLM&model_name=qwen:1.8b HTTP/1.1" 200 542
2024-03-24 19:40:08 [basehttp INFO] "GET /api/provider/model_list?provider=model_ollama_provider&model_type=LLM HTTP/1.1" 200 2538
2024-03-24 19:40:33 [basehttp INFO] "PUT /api/model/579cf4de-e995-11ee-a74e-0242ac100a08 HTTP/1.1" 200 303
2024-03-24 19:40:33 [basehttp INFO] "GET /api/model?name= HTTP/1.1" 200 305
2024-03-24 19:44:59.716 CST [29] LOG: checkpoint starting: time
2024-03-24 19:44:59.997 CST [29] LOG: checkpoint complete: wrote 2 buffers (0.0%); 0 WAL file(s) added, 0 removed, 0 recycled; write=0.132 s, sync=0.046 s, total=0.281 s; sync files=2, longest=0.031 s, average=0.023 s; distance=1 kB, estimate=3 kB
2024-03-24 20:36:51 [basehttp INFO] "GET /ui/assets/SyncWebDialog-5c10ee06.css HTTP/1.1" 200 280
2024-03-24 20:36:51 [basehttp INFO] "GET /ui/assets/index-bcb9597e.js HTTP/1.1" 200 4351
2024-03-24 20:36:51 [basehttp INFO] "GET /ui/assets/index-3090d140.css HTTP/1.1" 200 295
2024-03-24 20:36:51 [basehttp INFO] "GET /ui/assets/SyncWebDialog-1c3d5bfd.js HTTP/1.1" 200 2158
/usr/local/lib/python3.11/site-packages/django/db/models/base.py:364: RuntimeWarning: Model 'dataset.dynamics' was already registered. Reloading models is not advised as it can lead to inconsistencies, most notably with related models.
new_class._meta.apps.register_model(new_class._meta.app_label, new_class)
2024-03-24 20:36:51 [basehttp INFO] "GET /api/dataset/1/20 HTTP/1.1" 200 414
2024-03-24 20:36:51 [basehttp INFO] "GET /ui/assets/icon_document-2fa30876.svg HTTP/1.1" 200 1376
2024-03-24 20:36:53 [basehttp INFO] "GET /api/application/1/20 HTTP/1.1" 200 1988
2024-03-24 20:36:54 [basehttp INFO] "GET /ui/assets/index-b44a8f06.js HTTP/1.1" 200 10328
2024-03-24 20:36:54 [basehttp INFO] "GET /ui/assets/index-49976e2c.css HTTP/1.1" 200 405
2024-03-24 20:36:54 [basehttp INFO] "GET /api/application HTTP/1.1" 200 1937
2024-03-24 20:36:54 [basehttp INFO] "GET /api/application/fcbfad3a-e995-11ee-a74e-0242ac100a08/access_token HTTP/1.1" 200 223
2024-03-24 20:36:54 [basehttp INFO] "GET /api/application/fcbfad3a-e995-11ee-a74e-0242ac100a08 HTTP/1.1" 200 1996
2024-03-24 20:36:55 [basehttp INFO] "GET /ui/chat/34b6f605902a90af HTTP/1.1" 200 460
2024-03-24 20:36:55 [basehttp INFO] "GET /ui/assets/index-1b0dc722.js HTTP/1.1" 200 3259345
2024-03-24 20:36:55 [basehttp INFO] "GET /ui/assets/index-b342bee3.css HTTP/1.1" 200 433601
2024-03-24 20:37:02 [basehttp INFO] "GET /api/user_manage/1/20 HTTP/1.1" 200 369
2024-03-24 20:37:04 [basehttp INFO] "GET /api/provider HTTP/1.1" 200 17325
2024-03-24 20:37:04 [basehttp INFO] "GET /api/model?name= HTTP/1.1" 200 305
2024-03-24 20:37:06 [basehttp INFO] "GET /api/provider/model_type_list?provider=model_ollama_provider HTTP/1.1" 200 109
2024-03-24 20:37:06 [basehttp INFO] "GET /api/provider/model_form?provider=model_ollama_provider&model_type=LLM&model_name=qwen:1.8b HTTP/1.1" 200 542
2024-03-24 20:37:06 [basehttp INFO] "GET /api/provider/model_list?provider=model_ollama_provider&model_type=LLM HTTP/1.1" 200 2538
2024-03-24 20:37:21 [basehttp INFO] "GET /ui/assets/index-0531091d.js HTTP/1.1" 200 968
2024-03-24 20:37:21 [basehttp INFO] "GET /ui/assets/index-9da5a7b8.css HTTP/1.1" 200 805
2024-03-24 20:37:21 [basehttp INFO] "GET /ui/assets/icon_send-280a4ce1.svg HTTP/1.1" 200 945
2024-03-24 20:37:21 [basehttp INFO] "GET /ui/assets/icon_send_colorful-1f161525.svg HTTP/1.1" 200 1415
2024-03-24 20:37:21 [basehttp INFO] "POST /api/application/authentication HTTP/1.1" 200 403
2024-03-24 20:37:21 [basehttp INFO] "GET /api/application/profile HTTP/1.1" 200 508
2024-03-24 20:37:22 [basehttp INFO] "GET /ui/assets/icon_robot-c95ad06b.svg HTTP/1.1" 200 1337
2024-03-24 20:37:40 [basehttp INFO] "PUT /api/model/579cf4de-e995-11ee-a74e-0242ac100a08 HTTP/1.1" 200 303
2024-03-24 20:37:40 [basehttp INFO] "GET /api/model?name= HTTP/1.1" 200 305
2024-03-24 20:37:43 [basehttp INFO] "GET /ui/assets/user-icon-c413d294.svg HTTP/1.1" 200 1033
2024-03-24 20:37:43 [basehttp INFO] "GET /api/application/fcbfad3a-e995-11ee-a74e-0242ac100a08/chat/open HTTP/1.1" 200 88
/usr/local/lib/python3.11/site-packages/langchain_core/_api/deprecation.py:117: LangChainDeprecationWarning: The function `__call__` was deprecated in LangChain 0.1.7 and will be removed in 0.2.0. Use invoke instead.
warn_deprecated(
2024-03-24 20:37:44 [handle_exception ERROR] substring not found:Traceback (most recent call last):
File "/usr/local/lib/python3.11/site-packages/rest_framework/views.py", line 506, in dispatch
response = handler(request, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/maxkb/app/apps/common/auth/authentication.py", line 94, in run
return func(view, request, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/maxkb/app/apps/application/views/chat_views.py", line 83, in post
'client_type': request.auth.client_type}).chat()
^^^^^^
File "/opt/maxkb/app/apps/application/serializers/chat_message_serializers.py", line 191, in chat
pipline_message.run(params)
File "/opt/maxkb/app/apps/application/chat_pipeline/pipeline_manage.py", line 29, in run
step.run(self)
File "/opt/maxkb/app/apps/application/chat_pipeline/I_base_chat_pipeline.py", line 122, in run
self._run(manage)
File "/opt/maxkb/app/apps/application/chat_pipeline/step/reset_problem_step/i_reset_problem_step.py", line 37, in _run
padding_problem = self.execute(**self.context.get('step_args'))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/maxkb/app/apps/application/chat_pipeline/step/reset_problem_step/impl/base_reset_problem_step.py", line 32, in execute
padding_problem = response.content[response.content.index('<data>') + 6:response.content.index('</data>')]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: substring not found
2024-03-24 20:37:44 [basehttp INFO] "POST /api/application/chat_message/4f4b48ee-e9db-11ee-a973-0242ac100a06 HTTP/1.1" 200 61
(base) admin@ai-server:~$ curl http://172.16.10.2:11434/v1/chat/completions -d '{
"model": "qwen:1.8b",
"stream": false,
"messages": [
{ "role": "user", "content": "1+1=?" }
]
}'
{"id":"chatcmpl-401","object":"chat.completion","created":1711283990,"model":"qwen:1.8b","system_fingerprint":"fp_ollama","choices":[{"index":0,"message":{"role":"assistant","content":"1 + 1 = 2\n"},"finish_reason":"stop"}],"usage":{"prompt_tokens":9,"completion_tokens":10,"total_tokens":19}}
No response
v1.0.0 (build at 2024-04-10T19:09, commit: cbaf704)
本地搜索对英文大小写敏感,期望RAG过程应该是对大小写不敏感的
录入英文文档,对话中输入英文关键字即可100%复现
No response
No response
No response
版本号:v0.9.0 (build at 2024-03-21T18:39, commit: 96a0e4b)
尝试修改当前admin用户的密码,提示报错:admin(默认名称)的长度不足。具体截图。
用户管理界面,点击“编辑”。
保存新的邮箱信息。
默认的用户名是:admin
提示的长度是 5-20 个字符串。
校验时,最短字符串规则是:最小6个。
No response
版本号:v0.9.0 (build at 2024-03-21T18:39, commit: 96a0e4b)
我按照 使用手册 的方法添加自定义的Ollama ,在保存时报错,提示:校验失败,请检查 api_key secret_key 是否正确。
点击【添加模型】,选择供应商【 Ollama 】,直接进入下一步填写 Ollama 供应商的大模型表单。
点击【添加】后 校验不通过。
添加大模型到列表中。但是报错。
No response
经过测试,自定义域名是可正常访问。具体如下:
(base) root@ai-server:/var/lib/docker# curl http://172.16.10.10:11434/api/chat -d '{
"model": "qwen:1.8b",
"messages": [
{ "role": "user", "content": "1+1=?" }
]
}'
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.022455615Z","message":{"role":"assistant","content":"1"},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.05084234Z","message":{"role":"assistant","content":" +"},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.07471859Z","message":{"role":"assistant","content":" "},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.098813239Z","message":{"role":"assistant","content":"1"},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.122560097Z","message":{"role":"assistant","content":" ="},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.146415467Z","message":{"role":"assistant","content":" "},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.173886424Z","message":{"role":"assistant","content":"2"},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.226332598Z","message":{"role":"assistant","content":"\n"},"done":false}
{"model":"qwen:1.8b","created_at":"2024-03-23T15:42:52.250628975Z","message":{"role":"assistant","content":""},"done":true,"total_duration":331796019,"load_duration":124841,"prompt_eval_count":9,"prompt_eval_duration":103090000,"eval_count":10,"eval_duration":228125000}
V1.0.0
希望用户提问时命中人工维护过的分段内容可不走AI模型直接返回。
No response
No response
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.