# api-for-open-llm **Repository Path**: poe-ge/api-for-open-llm ## Basic Information - **Project Name**: api-for-open-llm - **Description**: jx - **Primary Language**: Python - **License**: Apache-2.0 - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 1 - **Created**: 2023-08-21 - **Last Updated**: 2023-08-21 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # API for Open LLMs

![llm.png](images/llm.png)
图片来自于论文: [A Survey of Large Language Models](https://arxiv.org/pdf/2303.18223.pdf)
## 📢 News + 【2023.08.10】 支持 [vLLM](https://github.com/vllm-project/vllm) 推理加速 ，使用方式见 [VLLM_SCRIPT](./docs/VLLM_SCRIPT.md) + 【2023.08.04】 增加通义千问模型工具调用示例，支持 **`function call`** 特性，调用方式参考 [邮件发送助手](./examples/email_sender.py)、[定积分计算器](./examples/quad_calculator.py)、[SQL查询](./examples/sql_querier.py) + 【2023.08.03】 支持 [qwen-7b-chat](https://github.com/QwenLM/Qwen-7B) 模型，启动方式见 [qwen-7b-chat](https://github.com/xusenlinzy/api-for-open-llm/blob/master/docs/SCRIPT.md#qwen-7b-chat) + 【2023.07.15】 支持 [starchat](https://huggingface.co/HuggingFaceH4/starchat-beta) 模型，启动方式见 [starchat](https://github.com/xusenlinzy/api-for-open-llm/blob/master/docs/SCRIPT.md#starchat) + 【2023.07.14】 支持在一个 **`API BASE`** 下调用多个模型 + ① 根据 [SCRIPT](docs/SCRIPT.md) 启动多个模型 + ② 修改 [chatgpt](./chatgpt.py) 中的 **`MODEL_LIST`**，将启动的模型加入进去 + ③ **`python chatgpt.py`** + 【2023.07.12】 支持 [baichuan-13b-chat](https://github.com/baichuan-inc/Baichuan-13B) 模型，启动方式见 [baichuan-13b-chat](https://github.com/xusenlinzy/api-for-open-llm/blob/master/docs/SCRIPT.md#baichuan-13b-chat) + 【2023.07.07】 支持 [InternLM](https://github.com/InternLM/InternLM) 模型，启动方式见 [internlm-chat-7b](https://github.com/xusenlinzy/api-for-open-llm/blob/master/docs/SCRIPT.md#internlm) + 【2023.06.26】 支持 [ChatGLM2-6b](https://github.com/THUDM/ChatGLM2-6B) 模型 + 【2023.06.12】 使用 [m3e](https://huggingface.co/moka-ai/m3e-base) 中文嵌入模型（在中文文本分类和文本检索上都优于 `openai-ada-002`） --- **主要特性**： 此项目为开源大模型的推理实现统一的后端接口，与 `OpenAI` 的响应保持一致，具有以下特性： + ✨ 以 `OpenAI ChatGPT API` 的方式调用各类开源大模型 + 🖨️ 支持流式响应，实现打印机效果 + 📖 实现文本嵌入模型，为文档知识问答提供支持 + 🦜️ 支持大规模语言模型开发工具 [`langchain` ](https://github.com/hwchase17/langchain) 的各类功能 + 🙌 只需要简单的修改环境变量即可将开源模型作为 `chatgpt` 的替代模型，为各类应用提供后端支持 + 🚀 支持加载经过自行训练过的 `lora` 模型 ## 🐼 模型 支持多种开源大模型 | Model | Backbone | #Params | Claimed language | checkpoint link | |:----------------------------------------------------------------------|:------------:|:--------:|:----------------:|:-----------------------------------------------------------------------------------------------------------:| | [qwen-7b-chat](https://github.com/QwenLM/Qwen-7B) | Qwen | 7B | en, zh | [Qwen/Qwen-7B-Chat](https://huggingface.co/baichuan-inc/Qwen/Qwen-7B-Chat) | | [baichuan-13b-chat](https://github.com/baichuan-inc/Baichuan-13B) | Baichuan | 13B | en, zh | [baichuan-inc/Baichuan-13B-Chat](https://huggingface.co/baichuan-inc/Baichuan-13B-Chat) | | [InternLM](https://github.com/InternLM/InternLM) | InternLM | 7B | en, zh | [internlm/internlm-chat-7b](https://huggingface.co/internlm/internlm-chat-7b) | | [ChatGLM2](https://github.com/THUDM/ChatGLM2-6B) | GLM | 6/130B | en, zh | [THUDM/chatglm2-6b](https://huggingface.co/THUDM/chatglm2-6b) | | [baichaun-7b](https://github.com/baichuan-inc/baichuan-7B) | Baichuan | 7B | en, zh | [baichuan-inc/baichuan-7B](https://huggingface.co/baichuan-inc/baichuan-7B) | | [Guanaco](https://github.com/artidoro/qlora/tree/main) | LLaMA | 7/33/65B | en | [timdettmers/guanaco-33b-merged](https://huggingface.co/timdettmers/guanaco-33b-merged) | | [YuLan-Chat](https://github.com/RUC-GSAI/YuLan-Chat) | LLaMA | 13/65B | en, zh | [RUCAIBox/YuLan-Chat-13b-delta](https://huggingface.co/RUCAIBox/YuLan-Chat-13b-delta) | | [TigerBot](https://github.com/TigerResearch/TigerBot) | BLOOMZ | 7/180B | en, zh | [TigerResearch/tigerbot-7b-sft](https://huggingface.co/TigerResearch/tigerbot-7b-sft) | | [OpenBuddy](https://github.com/OpenBuddy/OpenBuddy) | LLaMA、Falcon | 7B | multi | [OpenBuddy](https://huggingface.co/OpenBuddy) | | [MOSS](https://github.com/OpenLMLab/MOSS) | CodeGen | 16B | en, zh | [fnlp/moss-moon-003-sft-int4](https://huggingface.co/fnlp/moss-moon-003-sft-int4) | | [Phoenix](https://github.com/FreedomIntelligence/LLMZoo) | BLOOMZ | 7B | multi | [FreedomIntelligence/phoenix-inst-chat-7b](https://huggingface.co/FreedomIntelligence/phoenix-inst-chat-7b) | | [BAIZE](https://github.com/project-baize/baize-chatbot) | LLaMA | 7/13/30B | en | [project-baize/baize-lora-7B](https://huggingface.co/project-baize/baize-lora-7B) | | [Chinese-LLaMA-Alpaca](https://github.com/ymcui/Chinese-LLaMA-Alpaca) | LLaMA | 7/13B | en, zh | [ziqingyang/chinese-alpaca-plus-lora-7b](https://huggingface.co/ziqingyang/chinese-alpaca-plus-lora-7b) | | [BELLE](https://github.com/LianjiaTech/BELLE) | BLOOMZ | 7B | zh | [BelleGroup/BELLE-7B-2M](https://huggingface.co/BelleGroup/BELLE-7B-2M) | | [ChatGLM](https://github.com/THUDM/ChatGLM-6B) | GLM | 6B | en, zh | [THUDM/chatglm-6b](https://huggingface.co/THUDM/chatglm-6b) | ## 🐳 环境配置 ### docker启动（**推荐**） 构建镜像 ```shell docker build -f docker/Dockerfile -t llm-api:pytorch . ``` 启动容器 ```shell docker run -it -d --gpus all --ipc=host --net=host -p 80:80 --name=chatglm \ --ulimit memlock=-1 --ulimit stack=67108864 \ -v `pwd`:/workspace \ llm-api:pytorch \ python api/app.py \ --port 80 \ --allow-credentials \ --model_name chatglm \ --model_path THUDM/chatglm-6b \ --embedding_name moka-ai/m3e-base ``` 主要参数含义： + `model_name`: 模型名称，如`chatglm`、`phoenix`、`moss`等 + `model_path`: 开源大模型的文件所在路径 + `embedding_name`（可选项）: 嵌入模型的文件所在路径 更多模型的启动命令详见 [SCRIPT.md](docs/SCRIPT.md) ### 本地启动 安装 `pytorch` 环境 ```shell conda create -n pytorch python=3.8 conda activate pytorch conda install pytorch cudatoolkit -c pytorch ``` 安装依赖包 ```shell pip install -r requirements.txt ``` 启动后端 ```shell python api/app.py \ --port 80 \ --allow-credentials \ --model_path THUDM/chatglm-6b \ --embedding_name moka-ai/m3e-base ``` ## 🤖 使用方式 ### 环境变量 + `OPENAI_API_KEY`: 此处随意填一个字符串即可 + `OPENAI_API_BASE`: 后端启动的接口地址，如：http://192.168.0.xx:80/v1 ### [聊天界面](./applications) ```shell cd applications python web_demo.py ``` ![chat](images/chat.png) ### [openai-python](https://github.com/openai/openai-python)
👉 Chat Completions ```python import openai openai.api_base = "http://192.168.0.xx:80/v1" # Enter any non-empty API key to pass the client library's check. openai.api_key = "xxx" # Enter any non-empty model name to pass the client library's check. completion = openai.ChatCompletion.create( model="chatglm-6b", messages=[ {"role": "user", "content": "你好"}, ], stream=False, ) print(completion.choices[0].message.content) # 你好👋！我是人工智能助手 ChatGLM-6B，很高兴见到你，欢迎问我任何问题。 ```
👉 Completions ```python import openai openai.api_base = "http://192.168.0.xx:80/v1" # Enter any non-empty API key to pass the client library's check. openai.api_key = "xxx" # Enter any non-empty model name to pass the client library's check. completion = openai.Completion.create(prompt="你好", model="chatglm-6b") print(completion.choices[0].text) # 你好👋！我是人工智能助手 ChatGLM-6B，很高兴见到你，欢迎问我任何问题。 ```
👉 Embeddings ```python import openai openai.api_base = "http://192.168.0.xx:80/v1" # Enter any non-empty API key to pass the client library's check. openai.api_key = "xxx" # compute the embedding of the text embedding = openai.Embedding.create( input="什么是chatgpt？", model="text2vec-large-chinese" ) print(embedding['data'][0]['embedding']) ```
### [langchain](https://github.com/hwchase17/langchain)
👉 Chat Completions ```python import os os.environ["OPENAI_API_BASE"] = "http://192.168.0.xx:80/v1" os.environ["OPENAI_API_KEY"] = "xxx" from langchain.chat_models import ChatOpenAI from langchain.schema import HumanMessage chat = ChatOpenAI() print(chat([HumanMessage(content="你好")])) # content='你好👋！我是人工智能助手 ChatGLM-6B，很高兴见到你，欢迎问我任何问题。' additional_kwargs={} ```
👉 Completions ```python import os os.environ["OPENAI_API_BASE"] = "http://192.168.0.xx:80/v1" os.environ["OPENAI_API_KEY"] = "xxx" from langchain.llms import OpenAI llm = OpenAI() print(llm("你好")) # 你好👋！我是人工智能助手 ChatGLM-6B，很高兴见到你，欢迎问我任何问题。 ```
👉 Embeddings ```python import os os.environ["OPENAI_API_BASE"] = "http://192.168.0.xx:80/v1" os.environ["OPENAI_API_KEY"] = "xxx" from langchain.embeddings import OpenAIEmbeddings embeddings = OpenAIEmbeddings() query_result = embeddings.embed_query("什么是chatgpt？") print(query_result) ```
### 可接入的项目 **通过修改上面的 `OPENAI_API_BASE` 环境变量，大部分的 `chatgpt` 应用和前后端项目都可以无缝衔接！** + [ChatGPT-Next-Web: One-Click to deploy well-designed ChatGPT web UI on Vercel](https://github.com/Yidadaa/ChatGPT-Next-Web) ```shell docker run -d -p 3000:3000 \ -e OPENAI_API_KEY="sk-xxxx" \ -e BASE_URL="http://192.168.0.xx:80" \ yidadaa/chatgpt-next-web ``` ![web](images/web.png) + [dify: An easy-to-use LLMOps platform designed to empower more people to create sustainable, AI-native applications](https://github.com/langgenius/dify) ```shell # 在docker-compose.yml中的api和worker服务中添加以下环境变量 OPENAI_API_BASE: http://192.168.0.xx:80/v1 DISABLE_PROVIDER_CONFIG_VALIDATION: 'true' ``` ![dify](images/dify.png) ## 📜 License 此项目为 `Apache 2.0` 许可证授权，有关详细信息，请参阅 [LICENSE](LICENSE) 文件。 ## 🚧 References + [ChatGLM: An Open Bilingual Dialogue Language Model](https://github.com/THUDM/ChatGLM-6B) + [BLOOM: A 176B-Parameter Open-Access Multilingual Language Model](https://arxiv.org/abs/2211.05100) + [LLaMA: Open and Efficient Foundation Language Models](https://arxiv.org/abs/2302.13971v1) + [Efficient and Effective Text Encoding for Chinese LLaMA and Alpaca](https://github.com/ymcui/Chinese-LLaMA-Alpaca) + [Phoenix: Democratizing ChatGPT across Languages](https://github.com/FreedomIntelligence/LLMZoo) + [MOSS: An open-sourced plugin-augmented conversational language model](https://github.com/OpenLMLab/MOSS) + [FastChat: An open platform for training, serving, and evaluating large language model based chatbots](https://github.com/lm-sys/FastChat) + [LangChain: Building applications with LLMs through composability](https://github.com/hwchase17/langchain) + [ChuanhuChatgpt](https://github.com/GaiZhenbiao/ChuanhuChatGPT)