在 Hugging Face 上搭建 ChatGPT 聊天机器人

目录

• Hugging Face 上创建 ChatGPT Space
  • 克隆
  • 创建应用（聊天机器人）
    • chat.py
    • app.py
    • requirements.txt
  • 提交 Hugging Face
• 运行 ChatGPT
  • 演示
    • 设置 OPENAI_API_KEY
  • 日志（Logs）

Hugging Face 上创建 ChatGPT Space

克隆

git clone https://huggingface.co/spaces/wangjunjian/ChatGPT
cd ChatGPT

创建应用（聊天机器人）

chat.py

这里的 Conversation 记录了所有的对话消息，在提问前，会检查是否超过最大 token 数量，如果超过，会删除第一条与用户的对话消息，然后再提问。

import openai
import tiktoken


class Conversation:
    def __init__(self, prompt, model="gpt-3.5-turbo", temperature=0.8, max_tokens=250):
        self.prompt = prompt
        self.model = model
        self.temperature = temperature
        self.max_tokens = max_tokens

        self._init_messages()

    def _init_messages(self):
        self.messages = [{"role": "system", "content": self.prompt}]

    def reset(self):
        self._init_messages()

    def ask(self, question, pprint=True):
        self.messages.append({"role": "user", "content": question})

        if self.num_tokens(self.messages, self.model) >= self.max_tokens:
            if len(self.messages) > 3:
                self.messages = self.messages[:1] + self.messages[3:] # remove the first user message
            else:
                return "Error: max tokens exceeded."
    
        try:
            response = openai.ChatCompletion.create(
                model=self.model,
                messages=self.messages
            )   
        except Exception as e:
            return e

        if pprint:
            print(f"tiktoken: {self.num_tokens(self.messages, self.model)}\ntokens: {response['usage']}")
    
        assistant_message = response["choices"][0]["message"]["content"]
        self.messages.append({"role": "assistant", "content": assistant_message})

        return assistant_message
    
    def num_tokens(self, messages, model):
        """Returns the number of tokens used by a list of messages."""
        try:
            encoding = tiktoken.encoding_for_model(model)
        except KeyError:
            print("Warning: model not found. Using cl100k_base encoding.")
            encoding = tiktoken.get_encoding("cl100k_base")
        if model == "gpt-3.5-turbo":
            print("Warning: gpt-3.5-turbo may change over time. Returning num tokens assuming gpt-3.5-turbo-0301.")
            return self.num_tokens(messages, model="gpt-3.5-turbo-0301")
        elif model == "gpt-4":
            print("Warning: gpt-4 may change over time. Returning num tokens assuming gpt-4-0314.")
            return self.num_tokens(messages, model="gpt-4-0314")
        elif model == "gpt-3.5-turbo-0301":
            tokens_per_message = 4  # every message follows <|start|>{role/name}\n{content}<|end|>\n
            tokens_per_name = -1  # if there's a name, the role is omitted
        elif model == "gpt-4-0314":
            tokens_per_message = 3 
            tokens_per_name = 1 
        else:
            raise NotImplementedError(f"""num_tokens_from_messages() is not implemented for model {model}. See https://github.com/openai/openai-python/blob/main/chatml.md for information on how messages are converted to tokens.""")
        num_tokens = 0
        for message in messages:
            num_tokens += tokens_per_message
            for key, value in message.items():
                num_tokens += len(encoding.encode(value))
                if key == "name":
                    num_tokens += tokens_per_name
        num_tokens += 3  # every reply is primed with <|start|>assistant<|message|>
        return num_tokens

app.py

这里增加了 system 消息，让对话可以聚焦于某个主题，例如美食。对话前，用户可以在界面中设置自己的 OpenAI API Key，也可以 clone 为自己的 Space，然后通过 settings 的 Create Secret 来设置。

import os
import openai
import gradio as gr
from chat import Conversation

system_prompt_foodie = """您是一名美食家，帮助别人了解美食的问题。您的回答需要满足以下要求：
1. 回答要使用中文。
2. 回答要控制在100字以内。""" 
conv = Conversation(system_prompt_foodie, max_tokens=1024)

with gr.Blocks(title="ChatGPT 助手") as demo:
    openai_api_key = gr.Textbox(label="OpenAI API Key")
    chatbot = gr.Chatbot(elem_id="chatbot").style(height=500)
    msg = gr.Textbox(show_label=False, placeholder="Please enter your question...").style(container=False)
    clear = gr.Button("Clear")

    def set_openai_api_key(openai_api_key):
        openai.api_key = os.getenv("OPENAI_API_KEY")
        if openai_api_key:
            openai.api_key = openai_api_key
            
    openai_api_key.change(set_openai_api_key, openai_api_key)

    def ask(message, chat_history):
        if openai.api_key is None:
            chat_history.append((message, "Error: No OpenAI API Key found. Please enter your key in the cell above."))
        else:
            bot_message = conv.ask(message)
            chat_history.append((message, bot_message))

        return "", chat_history

    msg.submit(ask, [msg, chatbot], [msg, chatbot])
    clear.click(lambda: conv.reset(), None, chatbot, queue=False)

demo.launch()

requirements.txt

依赖的安装包

openai
tiktoken

提交 Hugging Face

git add app.py chat.py requirements.txt
git commit -m "Add application file"
git push

运行 ChatGPT

演示

设置 OPENAI_API_KEY

如果克隆了 Space，可以在 settings 中设置 OPENAI_API_KEY，这样就不用每次都输入了。

日志（Logs）

===== Build Queued at 2023-04-29 12:19:42 / Commit SHA: c0c29d2 =====

--> FROM docker.io/library/python:3.8.9@sha256:49d05fff9cb3b185b15ffd92d8e6bd61c20aa916133dca2e3dbe0215270faf53
DONE 0.0s

--> RUN --mount=target=requirements.txt,source=requirements.txt 	pip install --no-cache-dir -r requirements.txt
CACHED

--> RUN pip install --no-cache-dir pip==22.3.1 &&     pip install --no-cache-dir         datasets         "huggingface-hub>=0.12.1" "protobuf<4" "click<8.1"
CACHED

--> RUN useradd -m -u 1000 user
CACHED

--> WORKDIR /home/user/app
CACHED

--> RUN sed -i 's http://deb.debian.org http://cdn-aws.deb.debian.org g' /etc/apt/sources.list && sed -i 's http://archive.ubuntu.com http://us-east-1.ec2.archive.ubuntu.com g' /etc/apt/sources.list && sed -i '/security/d' /etc/apt/sources.list && apt-get update && apt-get install -y 	git 	git-lfs 	ffmpeg 	libsm6 	libxext6 	cmake 	libgl1-mesa-glx 	&& rm -rf /var/lib/apt/lists/* 	&& git lfs install
CACHED

--> RUN --mount=target=pre-requirements.txt,source=pre-requirements.txt 	pip install --no-cache-dir -r pre-requirements.txt
CACHED

--> RUN --mount=target=/root/packages.txt,source=packages.txt 	sed -i 's http://deb.debian.org http://cdn-aws.deb.debian.org g' /etc/apt/sources.list && sed -i 's http://archive.ubuntu.com http://us-east-1.ec2.archive.ubuntu.com g' /etc/apt/sources.list && sed -i '/security/d' /etc/apt/sources.list && apt-get update &&     xargs -r -a /root/packages.txt apt-get install -y     && rm -rf /var/lib/apt/lists/*
CACHED

--> COPY --link --chown=1000 --from=lfs /app /home/user/app
CACHED

--> RUN pip install --no-cache-dir         gradio==3.28.0
CACHED

--> COPY --link --chown=1000 ./ /home/user/app
DONE 0.0s

--> Pushing image
DONE 0.8s

--> Exporting cache
DONE 1.3s



===== Application Startup at 2023-04-29 13:41:27 =====

Running on local URL:  http://0.0.0.0:7860

To create a public link, set `share=True` in `launch()`.
Warning: gpt-3.5-turbo may change over time. Returning num tokens assuming gpt-3.5-turbo-0301.
Warning: gpt-3.5-turbo may change over time. Returning num tokens assuming gpt-3.5-turbo-0301.
tiktoken: 92
tokens: {
  "completion_tokens": 246,
  "prompt_tokens": 92,
  "total_tokens": 338
}
Warning: gpt-3.5-turbo may change over time. Returning num tokens assuming gpt-3.5-turbo-0301.
Warning: gpt-3.5-turbo may change over time. Returning num tokens assuming gpt-3.5-turbo-0301.
tiktoken: 358
tokens: {
  "completion_tokens": 351,
  "prompt_tokens": 358,
  "total_tokens": 709
}

RUN --mount=target=requirements.txt,source=requirements.txt 	pip install --no-cache-dir -r requirements.txt
--mount 选项来将本地的 requirements.txt 文件挂载到 Docker 容器中的目标路径。
--no-cache-dir 选项来告诉 pip 不要使用缓存，在安装依赖项时始终从远程源重新下载最新版本。这有助于确保容器中的依赖项始终是最新的。