From 486b30a8c56ab79c9520af7dd54b06f36849ca94 Mon Sep 17 00:00:00 2001 From: aries_ckt <916701291@qq.com> Date: Mon, 25 Sep 2023 21:09:17 +0800 Subject: [PATCH] doc:llm faq --- .../getting_started/faq/llm/llm_faq.po | 56 +++++++++++-------- pilot/server/llm_manage/api.py | 6 +- 2 files changed, 35 insertions(+), 27 deletions(-) diff --git a/docs/locales/zh_CN/LC_MESSAGES/getting_started/faq/llm/llm_faq.po b/docs/locales/zh_CN/LC_MESSAGES/getting_started/faq/llm/llm_faq.po index 2f127555b..8b588f5ed 100644 --- a/docs/locales/zh_CN/LC_MESSAGES/getting_started/faq/llm/llm_faq.po +++ b/docs/locales/zh_CN/LC_MESSAGES/getting_started/faq/llm/llm_faq.po @@ -8,7 +8,7 @@ msgid "" msgstr "" "Project-Id-Version: DB-GPT 👏👏 0.3.5\n" "Report-Msgid-Bugs-To: \n" -"POT-Creation-Date: 2023-09-14 14:35+0800\n" +"POT-Creation-Date: 2023-09-25 20:58+0800\n" "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n" "Last-Translator: FULL NAME \n" "Language: zh_CN\n" @@ -19,33 +19,33 @@ msgstr "" "Content-Transfer-Encoding: 8bit\n" "Generated-By: Babel 2.12.1\n" -#: ../../getting_started/faq/llm/llm_faq.md:1 73790502b62745ec88bbe9fe124254f0 +#: ../../getting_started/faq/llm/llm_faq.md:1 0d4fc79dbfce4f968ab310de12d69f3b msgid "LLM USE FAQ" msgstr "LLM模型使用FAQ" -#: ../../getting_started/faq/llm/llm_faq.md:3 473bdd77bbb242f497f514e6e63d0c5f +#: ../../getting_started/faq/llm/llm_faq.md:3 08873df3ef2741dca8916c4c0d503b4f msgid "Q1:how to use openai chatgpt service" msgstr "我怎么使用OPENAI服务" -#: ../../getting_started/faq/llm/llm_faq.md:4 6e073181e48e4604a301f3d7359c91ef +#: ../../getting_started/faq/llm/llm_faq.md:4 7741b098acd347659ccf663b5323666c msgid "change your LLM_MODEL" msgstr "通过在.env文件设置LLM_MODEL" -#: ../../getting_started/faq/llm/llm_faq.md:9 a88cd162dca448b198c0551a70e70da3 +#: ../../getting_started/faq/llm/llm_faq.md:9 018115ec074c48739b730310a8bafa44 msgid "set your OPENAPI KEY" msgstr "set your OPENAPI KEY" -#: ../../getting_started/faq/llm/llm_faq.md:16 ebaa67e9d31f4c70b4bccbd4394d1c27 +#: ../../getting_started/faq/llm/llm_faq.md:16 42408d9c11994a848da41c3ab87d7a78 msgid "make sure your openapi API_KEY is available" msgstr "确认openapi API_KEY是否可用" -#: ../../getting_started/faq/llm/llm_faq.md:18 8e88363a43b9460dae90a772360dcc5a +#: ../../getting_started/faq/llm/llm_faq.md:18 d9aedc07578d4562bad0ba1f130651de msgid "" "Q2 What difference between `python dbgpt_server --light` and `python " "dbgpt_server`" msgstr "Q2 `python dbgpt_server --light` 和 `python dbgpt_server`的区别是什么?" -#: ../../getting_started/faq/llm/llm_faq.md:21 1bbf3891883b43659b7ef39ce5e91918 +#: ../../getting_started/faq/llm/llm_faq.md:21 03c03fedaa2f4bfdaefb42fd4164c902 msgid "" "`python dbgpt_server --light` dbgpt_server does not start the llm " "service. Users can deploy the llm service separately by using `python " @@ -57,54 +57,54 @@ msgstr "" "用户可以通过`python " "llmserver`单独部署模型服务，dbgpt_server通过LLM_SERVER环境变量来访问模型服务。目的是为了可以将dbgpt后台服务和大模型服务分离部署。" -#: ../../getting_started/faq/llm/llm_faq.md:23 96a6b6be655c4f85a7c18e813f67517e +#: ../../getting_started/faq/llm/llm_faq.md:23 61354a0859284346adc3e07c820aa61a msgid "" "`python dbgpt_server` dbgpt_server service and the llm service are " "deployed on the same instance. when dbgpt_server starts the service, it " "also starts the llm service at the same time." msgstr "`python dbgpt_server` 是将后台服务和模型服务部署在同一台实例上.dbgpt_server在启动服务的时候同时开启模型服务." -#: ../../getting_started/faq/llm/llm_faq.md:27 8a0138f4ceab476a97f112776669c7ca +#: ../../getting_started/faq/llm/llm_faq.md:27 41ee95bf0b224be995f7530d0b67f712 #, fuzzy msgid "Q3 how to use MultiGPUs" msgstr "Q2 怎么使用 MultiGPUs" -#: ../../getting_started/faq/llm/llm_faq.md:29 6b2f25a5a2b243f78c2f96e3b045bf97 +#: ../../getting_started/faq/llm/llm_faq.md:29 7fce22f0327646399b98b0e20574a2fd msgid "" "DB-GPT will use all available gpu by default. And you can modify the " "setting `CUDA_VISIBLE_DEVICES=0,1` in `.env` file to use the specific gpu" " IDs." msgstr "DB-GPT默认加载可利用的gpu，你也可以通过修改 在`.env`文件 `CUDA_VISIBLE_DEVICES=0,1`来指定gpu IDs" -#: ../../getting_started/faq/llm/llm_faq.md:32 2adf75ffb0ab451999d2f446389eea6c +#: ../../getting_started/faq/llm/llm_faq.md:32 3f4eb824dc924d7ca309dc5057f8360a msgid "" "Optionally, you can also specify the gpu ID to use before the starting " "command, as shown below:" msgstr "你也可以指定gpu ID启动" -#: ../../getting_started/faq/llm/llm_faq.md:42 793d6d8503b74323b4997cf2981cc098 +#: ../../getting_started/faq/llm/llm_faq.md:42 a77d72f91b864d0aac344b317c100950 msgid "" "You can modify the setting `MAX_GPU_MEMORY=xxGib` in `.env` file to " "configure the maximum memory used by each GPU." msgstr "同时你可以通过在.env文件设置`MAX_GPU_MEMORY=xxGib`修改每个GPU的最大使用内存" -#: ../../getting_started/faq/llm/llm_faq.md:44 bdfc8eb5bc89460ea3979f61b8aeca7f +#: ../../getting_started/faq/llm/llm_faq.md:44 b3bb92777a1244d5967a4308d14722fc #, fuzzy msgid "Q4 Not Enough Memory" msgstr "Q3 机器显存不够 " -#: ../../getting_started/faq/llm/llm_faq.md:46 e0e60f0263d34eec818b72c38d214b8f +#: ../../getting_started/faq/llm/llm_faq.md:46 c3976d81aafa4c6081e37c0d0a115d96 msgid "DB-GPT supported 8-bit quantization and 4-bit quantization." msgstr "DB-GPT 支持 8-bit quantization 和 4-bit quantization." -#: ../../getting_started/faq/llm/llm_faq.md:48 98c954d9fcf449f4b47610fc96091c4f +#: ../../getting_started/faq/llm/llm_faq.md:48 93ade142f949449d8f54c0b6d8c8d261 msgid "" "You can modify the setting `QUANTIZE_8bit=True` or `QUANTIZE_4bit=True` " "in `.env` file to use quantization(8-bit quantization is enabled by " "default)." msgstr "你可以通过在.env文件设置`QUANTIZE_8bit=True` or `QUANTIZE_4bit=True`" -#: ../../getting_started/faq/llm/llm_faq.md:50 2568b441f7e54654b405c7791f08036a +#: ../../getting_started/faq/llm/llm_faq.md:50 be2573907d624ebf8c901301f938577b msgid "" "Llama-2-70b with 8-bit quantization can run with 80 GB of VRAM, and 4-bit" " quantization can run with 48 GB of VRAM." @@ -112,41 +112,49 @@ msgstr "" "Llama-2-70b with 8-bit quantization 可以运行在 80 GB VRAM机器， 4-bit " "quantization可以运行在 48 GB VRAM" -#: ../../getting_started/faq/llm/llm_faq.md:52 f8d1e4312f9743c7b03820b4a8dbf992 +#: ../../getting_started/faq/llm/llm_faq.md:52 c084d4624e794f7e8ceebadb6f260b49 msgid "" "Note: you need to install the latest dependencies according to " "[requirements.txt](https://github.com/eosphoros-ai/DB-" "GPT/blob/main/requirements.txt)." msgstr "" -#: ../../getting_started/faq/llm/llm_faq.md:54 5fe0d9ced7e848799f4d7bce92a5c130 +#: ../../getting_started/faq/llm/llm_faq.md:54 559bcd62af7340f79f5eca817187e13e #, fuzzy msgid "Q5 How to Add LLM Service dynamic local mode" msgstr "Q5 怎样动态新增模型服务" -#: ../../getting_started/faq/llm/llm_faq.md:56 fd921148e3e547beb6c74035a6b6a8b0 +#: ../../getting_started/faq/llm/llm_faq.md:56 e47101d7d47e486e8572f6acd609fa92 msgid "" "Now DB-GPT through multi-llm service switch, so how to add llm service " "dynamic," msgstr "DB-GPT支持多个模型服务切换, 怎样添加一个模型服务呢" -#: ../../getting_started/faq/llm/llm_faq.md:67 5fe0d9ced7e848799f4d7bce92a5c130 +#: ../../getting_started/faq/llm/llm_faq.md:67 5710dd9bf8f54bd388354079b29acdd2 #, fuzzy msgid "Q6 How to Add LLM Service dynamic in remote mode" msgstr "Q5 怎样动态新增模型服务" -#: ../../getting_started/faq/llm/llm_faq.md:68 bd29cd6d29a64908af15b391d73ea82a +#: ../../getting_started/faq/llm/llm_faq.md:68 9c9311d6daad402a8e0748f00e69e8cf msgid "" "If you deploy llm service in remote machine instance, and you want to " "add model service to dbgpt server to manage" msgstr "如果你想在远程机器实例部署大模型服务并添加到本地dbgpt_server进行管理" -#: ../../getting_started/faq/llm/llm_faq.md:70 ace16dfc4326431dbe4a9a32e4a83ba4 +#: ../../getting_started/faq/llm/llm_faq.md:70 3ec1565e74384beab23df9d8d4a19a39 msgid "use dbgpt start worker and set --controller_addr." msgstr "使用1`dbgpt start worker`命令并设置注册地址--controller_addr" -#: ../../getting_started/faq/llm/llm_faq.md:81 f8c024339da447ce8160a4eb9f87c125 +#: ../../getting_started/faq/llm/llm_faq.md:80 e2b8a9119f7843beb787d021c973eea4 #, fuzzy msgid "Q7 dbgpt command not found" msgstr "Q6 dbgpt command not found" +#: ../../getting_started/faq/llm/llm_faq.md:86 257ae9c462cd4a9abe7d2ff00f6bc891 +msgid "" +"Q8 When starting the worker_manager on a cloud server and registering it " +"with the controller, it is noticed that the worker's exposed IP is a " +"private IP instead of a public IP, which leads to the inability to access" +" the service." +msgstr "云服务器启动worker_manager注册到controller时，发现worker暴露的ip是私网ip, 没有以公网ip暴露，导致服务访问不到" + diff --git a/pilot/server/llm_manage/api.py b/pilot/server/llm_manage/api.py index 04e90feb0..1ef62d794 100644 --- a/pilot/server/llm_manage/api.py +++ b/pilot/server/llm_manage/api.py @@ -66,15 +66,15 @@ async def model_list(): last_heartbeat=model.last_heartbeat, prompt_template=model.prompt_template, ) - response.manager_host = model.host if manager_map[model.host] else None + response.manager_host = model.host if manager_map.get(model.host) else None response.manager_port = ( - manager_map[model.host].port if manager_map[model.host] else None + manager_map[model.host].port if manager_map.get(model.host) else None ) responses.append(response) return Result.succ(responses) except Exception as e: - return Result.faild(code="E000X", msg=f"space list error {e}") + return Result.faild(code="E000X", msg=f"model list error {e}") @router.post("/v1/worker/model/stop")