文章目录
*
– 更新时间
– 环境
– 流程介绍
– 代码（onnx，darknet）
– 代码（pytorch）
– 参考文章
更新时间
最近更新： 2022-09
环境
Python 3.8
opencv 4.5.2.54
onnxruntime 1.10.0
pytorch 1.10.2
cryptography 3.1.1
流程介绍
加密：
- 模型以二进制存储
- 以二进制形式读取
- 用cryptography对文件加密
- 保存加密后文件
解密：
- 读取加密文件
- 使用cryptography解密
- 转换为框架可读数据
代码（onnx，darknet）
from cryptography.fernet import Fernet
key = Fernet.generate_key()
print(key)
f = Fernet(key)
model_file = ['./yolov4-tiny.weights','./yolov4-tiny.cfg','./yolov4-tiny.onnx']
new_model_file = ['./yolov4-tiny.dll','./yolov4-tiny-cfg.dll','./yolov4-tiny-onnx.dll']
for i in range(2):
with open(new_model_file[i],'wb') as ew:
content = open(model_file[i],'rb').read()
encrypted_content = f.encrypt(content)
ew.write(encrypted_content)
import cv2
conf_file = open(new_model_file[1],'rb').read()
conf_file = f.decrypt(conf_file)
conf_file = bytearray(conf_file)
weight_flie = open(new_model_file[0],'rb').read()
weight_flie = f.decrypt(weight_flie)
weight_flie = bytearray(weight_flie)
net = cv2.dnn.readNetFromDarknet(conf_file,weight_flie)
print(net)
import onnxruntime
onnx_file = open(new_model_file[2],'rb').read()
onnx_file = f.decrypt(onnx_file)
session = onnxruntime.InferenceSession(onnx_file)
print(session)
代码（pytorch）
from cryptography.fernet import Fernet
key = Fernet.generate_key()
print(key)
f = Fernet(key)
import io
import torch
model = torch.load('./model.pth',map_location='cpu')
byte = io.BytesIO()
torch.save(model, byte)
byte.seek(0)
pth_bytes = byte.read()
new_model = f.encrypt(pth_bytes)
with open('new_model.pth', 'wb') as fw:
fw.write(new_model)
with open('new_model.pth', 'rb') as fr:
new_model = fr.read()
new_model = f.decrypt(new_model)
model_byte = io.BytesIO(new_model)
model_byte.seek(0)
model = torch.load(model_byte)
print(model)
参考文章
Pytorch模型加密的方法
神经网络如何加密（Python）
Original: https://blog.csdn.net/djj199301111/article/details/126708162
Author: 炼丹去了
Title: 深度学习模型加密（pytorch，onnx，darknet）
原创文章受到原创版权保护。转载请注明出处：https://www.johngo689.com/708635/
转载文章受原作者版权保护。转载请注明原作者出处！