当前位置：首页 > news >正文

在魔塔社区搭建通义千问-7B（Qwen-7B）流程

news 2025/8/25 0:28:19

复制以下语句

python3 -m venv myvenvsource myvenv/bin/activatepip install modelscope
pip install transformers_stream_generator
pip install transformers
pip install tiktoken
pip install accelerate
pip install bitsandbytestouch run.py
vi run.py

复制下面代码粘贴到 run.py

import os
import platform
from modelscope import AutoModelForCausalLM, AutoTokenizer, GenerationConfigmodel_id = 'qwen/Qwen-7B-Chat'
revision = 'v1.0.1'
tokenizer = AutoTokenizer.from_pretrained(model_id, revision=revision, trust_remote_code=True)
# use fp16
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", revision=revision, trust_remote_code=True, fp16=True).eval()
model.generation_config = GenerationConfig.from_pretrained(model_id,trust_remote_code=True)  # 可指定不同的生成长度、top_p等相关超参stop_stream = Falsedef clear_screen():if platform.system() == "Windows":os.system("cls")else:os.system("clear")def print_history(history):for pair in history:print(f"\nUser：{pair[0]}\nQwen-7B：{pair[1]}")def main():history, response = [], ''global stop_streamclear_screen()print("欢迎使用 Qwen-7B 模型，输入内容即可进行对话，clear 清空对话历史，stop 终止程序")while True:query = input("\nUser：")if query.strip() == "stop":breakif query.strip() == "clear":history = []clear_screen()print("欢迎使用 Qwen-7B 模型，输入内容即可进行对话，clear 清空对话历史，stop 终止程序")continuefor response in model.chat(tokenizer, query, history=history, stream=True):if stop_stream:stop_stream = Falsebreakelse:clear_screen()print_history(history)print(f"\nUser: {query}")print("\nQwen-7B：", end="")print(response)history.append((query, response))if __name__ == "__main__":main()

按下以下按键: ESC键 :wq 回车
就保存好了

然后就运行