XXXX 本地模型替换为 两家 API

 llamaindex_internlm.py 

原代码

from llama_index.llms.huggingface import HuggingFaceLLM
from llama_index.core.llms import ChatMessage
llm = HuggingFaceLLM(
    model_name="/root/model/internlm2-chat-1_8b",
    tokenizer_name="/root/model/internlm2-chat-1_8b",
    model_kwargs={"trust_remote_code":True},
    tokenizer_kwargs={"trust_remote_code":True}
)

rsp = llm.chat(messages=[ChatMessage(content="xtuner是什么？")])
print(rsp)

替换后

硅基API

from llama_index.core.llms import ChatMessage
import requests


url = "https://api.siliconflow.cn/v1/chat/completions"

messages = [ChatMessage(content="xtuner是什么？")]

payload = {
    "model": "internlm/internlm2_5-7b-chat",
    "messages": [{"role": "user", "content": msg.content} for msg in messages]
}

headers = {
    "Authorization": "Bearer sk-XXXXXXXXXXXXXXXXXX",  # 替换为API key
    "Content-Type": "application/json"
}


response = requests.post(url, json=payload, headers=headers)

response_json = response.json()

assistant_message = response_json['choices'][0]['message']['content']
print(assistant_message)

运行结果

浦语API

from llama_index.core.llms import ChatMessage
import requests

# API endpoint for chat completions
url = "https://internlm-chat.intern-ai.org.cn/puyu/api/v1/chat/completions"

# Prepare the message using ChatMessage
messages = [ChatMessage(content="xtuner是什么？")]

# Prepare the payload with the model and formatted messages
payload = {
    "model": "internlm2.5-latest",
    "messages": [{"role": "user", "content": msg.content} for msg in messages],
    "n": 1,
    "temperature": 0.8,
    "top_p": 0.9
}

# Specify the authorization token and content type in the headers
headers = {
    "Authorization": "Bearer eyJ0eXBlIjoiSldUIiwiYWxnIjoiSFM1MTXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",  # 替换为你的实际token
    "Content-Type": "application/json"
}

# Send a POST request to the API
response = requests.post(url, json=payload, headers=headers)

# Parse the response JSON
response_json = response.json()

# Extract and print the content from the assistant's message
assistant_message = response_json['choices'][0]['message']['content']
print(assistant_message)

运行结果

llamaindex_RAG.py

原代码

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings

from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.llms.huggingface import HuggingFaceLLM

#初始化一个HuggingFaceEmbedding对象，用于将文本转换为向量表示
embed_model = HuggingFaceEmbedding(
#指定了一个预训练的sentence-transformer模型的路径
    model_name="/root/model/sentence-transformer"
)
#将创建的嵌入模型赋值给全局设置的embed_model属性，
#这样在后续的索引构建过程中就会使用这个模型。
Settings.embed_model = embed_model

llm = HuggingFaceLLM(
    model_name="/root/model/internlm2-chat-1_8b",
    tokenizer_name="/root/model/internlm2-chat-1_8b",
    model_kwargs={"trust_remote_code":True},
    tokenizer_kwargs={"trust_remote_code":True}
)
#设置全局的llm属性，这样在索引查询时会使用这个模型。
Settings.llm = llm

#从指定目录读取所有文档，并加载数据到内存中
documents = SimpleDirectoryReader("/root/llamaindex_demo/data").load_data()
#创建一个VectorStoreIndex，并使用之前加载的文档来构建索引。
# 此索引将文档转换为向量，并存储这些向量以便于快速检索。
index = VectorStoreIndex.from_documents(documents)
# 创建一个查询引擎，这个引擎可以接收查询并返回相关文档的响应。
query_engine = index.as_query_engine()
response = query_engine.query("xtuner是什么?")

print(response)

修改后：

硅基API

import requests
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core.settings import Settings

# 禁用全局LLM设置中的OpenAI，不然会报错
Settings.llm = None

# 初始化嵌入模型
embed_model = HuggingFaceEmbedding(
    model_name="/root/model/sentence-transformer"
)

# 从指定目录读取所有文档，并加载数据到内存中
documents = SimpleDirectoryReader("/root/llamaindex_demo/data").load_data()

# 创建一个VectorStoreIndex，指定使用自己的嵌入模型
index = VectorStoreIndex.from_documents(documents, embed_model=embed_model)

# 创建一个查询引擎用于本地文档查询，并禁用LLM
query_engine = index.as_query_engine(llm=None)

# 准备查询的问题
question = "xtuner是什么?"
local_response = query_engine.query(question)

# 将响应转变为字符串形式
local_response_str = str(local_response) 

# print("本地查询结果:", local_response_str)

# 使用自定义API进行外部查询
url = "https://api.siliconflow.cn/v1/chat/completions"
payload = {
    "model": "internlm/internlm2_5-7b-chat",
    "messages": [{"role": "user", "content": "对以下问题和片段进行总结"+ question + " " + local_response_str}]
}
headers = {
    "Authorization": "Bearer sk-fzmebtdhrqnhnsnnXXXXXXXXXXXXXXXXX",  # 使用你的实际API令牌代替
    "Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)

# 打印原始响应，以便调试
# print("原始API响应:", response.text)

# 解析响应并提取content字段
try:
    response_json = response.json()
    content = response_json['choices'][0]['message']['content']
    print(content)
except Exception as e:
    print(f"Error: {str(e)}")

运行结果：

浦语API

import requests
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from llama_index.core.settings import Settings
import json

# 禁用全局LLM设置中的OpenAI，不然会报错
Settings.llm = None

# 初始化嵌入模型
embed_model = HuggingFaceEmbedding(
    model_name="/root/model/sentence-transformer"
)

# 从指定目录读取所有文档，并加载数据到内存中
documents = SimpleDirectoryReader("/root/llamaindex_demo/data").load_data()

# 创建一个VectorStoreIndex，指定使用自己的嵌入模型
index = VectorStoreIndex.from_documents(documents, embed_model=embed_model)

# 创建一个查询引擎用于本地文档查询，并禁用LLM
query_engine = index.as_query_engine(llm=None)

# 准备查询的问题
question = "xtuner是什么?"
local_response = query_engine.query(question)

# 将响应转变为字符串形式
local_response_str = str(local_response)

# print("本地查询结果:", local_response_str)

# 使用自定义API进行外部查询
url = "https://internlm-chat.intern-ai.org.cn/puyu/api/v1/chat/completions"
payload = {
    "model": "internlm2.5-latest",
    "messages": [{"role": "user", "content": "对以下问题和片段进行总结"+ question + " " + local_response_str}],
    "n": 1,
    "temperature": 0.8,
    "top_p": 0.9
}
headers = {
    "Authorization": "Bearer eyJ0eXBlIjoiSldUIiwiYWxnIjoiSFM1MTIifQXXXXXXXXXXXX",  # 使用你的实际API令牌代替
    "Content-Type": "application/json"
}
response = requests.post(url, data=json.dumps(payload), headers=headers)

# 打印原始响应，以便调试
# print(response.text)

# 解析JSON响应并直接提取content
try:
    response_json = response.json()
    content = response_json['choices'][0]['message']['content']
    print(content)
except Exception as e:
    print(f"Error extracting content: {str(e)}")

运行结果：