Ubuntu20.04部署Qwen2.openvino流程_怎么在乌班图虚拟机上部署qwen大模型

下载代码

里面包含依赖

git clone https://github.com/OpenVINO-dev-contest/Qwen2.openvino.git
1

python环境配置

创建虚拟环境

conda create -name qwen2openvino python==3.10
conda activate qwen2openvino
1
2

安装依赖

pip install wheel setuptools
pip install -r requirements.txt
1
2

配置镜像服务器来加快下载速度

export HF_ENDPOINT=https://hf-mirror.com
1

下载模型

huggingface-cli download --resume-download --local-dir-use-symlinks False Qwen/Qwen1.5-0.5B-Chat --local-dir ./Qwen1.5-0.5B-Chat
1

转换模型

python convert.py --model_id ./Qwen1.5-0.5B-Chat --precision int4 --output ./Qwen1.5-0.5B-Chat-ov
1

在chat.py的前面加上，防止出现waring

import os
os.environ["TOKENIZERS_PARALLELISM"] = "false"
1
2

【问题解决】module ‘triton.compiler‘ has no attribute ‘OutOfResources‘

[已解決] huggingface/tokenizers: The current process just got forked. after parallelism has already been used. Disabling parallelism to avoid deadlocks

运行chat.py

python chat.py --model_path ./Qwen1.5-0.5B-Chat-ov --max_sequence_length 4096 --device GPU
1

参考文章

(Qwen)通义千问大模型安装部署教程2024最新