本地运行chatglm3-6b 和 ChatPromptTemplate的结合使用

import gradio
from transformers import AutoTokenizer, AutoModel
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_community.llms import HuggingFacePipeline
import time

def greet(name):
    response = chain.invoke({"user_input": name})
    return response

model = HuggingFacePipeline.from_model_id(
    model_id="THUDM/chatglm3-6b",
    task="text-generation",
    device=0,
    model_kwargs={"trust_remote_code":True},
    pipeline_kwargs={"max_new_tokens": 5000},
)

prompt = ChatPromptTemplate.from_template("告诉我关于{user_input}的经济发展情况，不多于200个字")
output_parser = StrOutputParser()
chain = prompt | model | output_parser
demo = gradio.Interface(fn=greet, inputs="text", outputs="text")
demo.launch() 

import gradio
from langchain_core.prompts import ChatPromptTemplate  
from langchain.prompts import PromptTemplate 
from langchain_core.output_parsers import StrOutputParser
from langchain_community.llms import HuggingFacePipeline
from langchain.prompts import HumanMessagePromptTemplate
import time

# from ChatGLM_new import zhipu_llm
# model  = zhipu_llm 


model = HuggingFacePipeline.from_model_id(
    model_id="THUDM/chatglm3-6b",
    task="text-generation",
    verbose=True,
    device=0,
    model_kwargs={"trust_remote_code":True},
    pipeline_kwargs={"max_new_tokens": 5000},
)

prompt = ChatPromptTemplate.from_messages([
                # ("system", "记住：对所有问题你只回答下面的4个字：我不知道，"),
                # ("human", "Hello, how are you doing?"),
                # ("ai", "I'm doing well, thanks!"),
                ("human", "告诉我关于{user_input}的经济发展情况，不多于200个字"),
            ])

prompt = ChatPromptTemplate.from_messages([
     HumanMessagePromptTemplate.from_template("告诉我关于{user_input}的经济发展情况，不多于200个字"),
            ])

output_parser = StrOutputParser()
chain = prompt | model | output_parser
def greet(name):
    response = chain.invoke({"user_input": name})
    return response
demo = gradio.Interface(fn=greet, inputs="text", outputs="text")
demo.launch()