环境配置

创建conda环境

1
conda create -n insightface python=3.8

激活环境

1
conda activate insightface

配置环境

1
2
3
4
5
conda install pytorch==1.10.0 torchvision==0.11.0 torchaudio==0.10.0 -c pytorch
conda install cudatoolkit=11.3
conda install cudnn==8.2.1
pip install onnxruntime-gpu==1.14.1
pip install opencv-contrib-python

注意安装onnxruntime-gpu记得去查询CUDAcuDNN的版本对应关系NVIDIA - CUDA | onnxruntime

环境测试

创建文件test.py

1
2
3
4
5
6
7
8
9
import onnxruntime as ort

# 检查是否有CUDA加速
available_providers = ort.get_available_providers()
print(f"Available providers: {available_providers}")

# 检查当前设备是GPU还是CPU
device = ort.get_device()
print(f"Current device: {device}")

查看运行结果(如果Current device: GPU,则表示正确)

image-20250428194655440

创建文件main.py

1
2
3
4
5
6
import insightface


app = insightface.app.FaceAnalysis(providers=['CUDAExecutionProvider'])
app.prepare(ctx_id=0)
print("insightface 初始化成功")
可能会遇到的问题

image-20250428194842995

修改代码为

1
2
3
4
5
6
7
8
9
# 在任何库导入前设置
import os

os.environ['NO_ALBUMENTATIONS_UPDATE'] = '1'
import insightface

app = insightface.app.FaceAnalysis(providers=['CUDAExecutionProvider'])
app.prepare(ctx_id=0)
print("insightface 初始化成功")

image-20250428195113745

配置成功